I thought I would focus this issue on articles about AI writing that have caught my attention, partly because I’ve recently had the pleasure of being on a couple of panels that were largely about this, including TORQUE in Bruges, on AI in scientific publishing, and the ISMTE conference, on AI safeguards, and partly because my views on AI writing are changing as I notice what it is doing to my writing and thinking.

AI disclosures and writing policies:

I think it’s important that a journal’s AI policies reflect the opinions of the communities who read and write for them and that we don’t have a one-size-fits-all approach.

• Philosophy & Public Affairs [tweet thread] explains their policy and why they want no AI-authored writing. “Academic journals serve at least two functions: the promulgation of new knowledge, and identification and credentialing of talented researchers... Submitting AI-authored essays makes the task of identifying talented researchers harder.”

• At the other end of the spectrum, Elsevier’s new Generative AI policies for journals are very permissive and allow AI use for literature reviews, identifying research gaps, generating ideas, structuring manuscripts, improving writing, checking manuscripts, and creating some explanatory images. See their FAQ for details.

• Florent Daudens on Stanley Druckenmiller's admission that his WSJ op-ed was written with AI: Who gets to use AI to write? What a billionaire, a Harvard professor, ghostwriters, and a few research papers tell us about the increasingly messy rules of authorship. "For the WSJ's opinion section, an AI-assisted op-ed can still be authentic if the underlying argument belongs to the person whose name appears on it. For the FT, the process of producing the words is also part of the editorial contract."
  
  A year ago I would have been in the WSJ camp, but now I'm veering towards the FT's position, especially for editorial content, for two reasons. The first is that I think AI writing patterns are becoming stronger and are very hard to get rid of, so I find myself dismissing writing as AI slop even if the ideas are good. Secondly, unless you are really clear about your view (which I think Stanley Druckenmiller would be), the AI anchors what you write. Even if you're just using it to take voice notes, it tends to give these a spin, which can subtly change your thinking. [I'm assuming AI is being used collaboratively and that the user is giving it their own thoughts and iterating on them, rather than asking it to write the whole thing.]

• I like Clay’s (a go-to-market automation platform) AI writing policy. It’s a simple approach, but requires a lot of user input:

  • You must stand behind every idea and sentence

  • Writing is thinking [I’m less certain this is true for all of us]

  • More time should be spent writing a document than consuming it

  • Longer is not better

• Scholarly Kitchen asks if Research Needs a Shared Standard for AI Disclosure? Whilst a neat and tidy AI disclosure taxonomy appeals to me, I don’t think this would be as stable or straightforward as CRediT contributor roles. I also suspect it might date badly and need regular updating.

• Claude will watermark text to comply with the EU AI Act. In this case, watermarking is done via word selection - see this accessible explainer about the technicalities of how this works. [I think that if you read enough unedited Claude-written text, you probably don’t need a watermark to identify Claude writing.] The Claude team note: “Can’t someone just edit the text to get around the watermarking? To some extent, yes. Light editing probably won’t remove the watermark completely; a complete rewrite where every word is replaced will.” I am uncertain whether this kind of watermarking has a lot of real-world value in scholarly publishing. I would have thought anyone creating papermill-type content isn’t going to be using Claude or can easily work around this.

• If you’re interested in testing how easy it is to fool Pangram with academic writing, here’s a simple exercise that takes around 15 minutes and follows the same passage as it is generated by AI, processed by a humaniser, and then copy-edited by another AI tool. Pangram 4 is harder to fool than earlier versions, but not that hard to fool:

🚀 Launches:

• Digital Science (Overleaf) has launched Papers AI, which says it “brings your writing, data, references and code into the same flow, so you can follow the thought without rebuilding the context.” Mark Hahnel’s write-up here.

• OpenAI is sunsetting its LaTeX document creation tool Prism, and OpenAI for Science, which was led by Kevin Weil, is being ‘decentralised’ [H/T: Ian Mulvaney]. Slightly off topic, but Kevin Weil is now seeking $150 million in funding for a new company to gather scientific data for AI models.

📊 Data:

• Caleb Robinson and Isaac Corley put together a Q&A from the slop trenches, including because it contains real data about what people are seeing: “We reviewed 22 ML conference submissions this summer. Fifteen had fabricated citations, hallucinated authors, or clear LLM slop, so we complain about that a bit, and release the paper references audit we now run.”

• Mahmud Omar asks: Are we losing our maybes? “Hedges fell 6.1% per 1,000 words. Boosters rose 31.4%. The hedge-to-booster ratio fell 28.6% (as you can see in the figure below).”

  

Reflections on AI writing:

• Dr Benn Gooch’s nuanced and interesting comments on abandoning the use of AI scribes after 18 months are worth a read. “I felt myself becoming a passive observer in encounters where I had previously been an active architect. I felt my clinical memory, my narrative identity, and my sense of connection to my patients beginning to erode at the edges.”

• Scott Bouchard at Silverchair posted What I got wrong (about AI) this year. “The enterprise and institutional deployments, where the access controls and privacy protections live, are not where most of the work is happening.” The thing I got most wrong is that I thought AI writing would be a lot better and largely undetectable by now. I’ve tried various AI-supported writing systems such as the “feed it your writing and build a voice profile” approach, various skills such as Mike Caulfield’s Jamesian skill, and an AI-tells removal skill similar to Dustin Smith’s Clean Prose skill, but on the whole, none of them seem to work very well, and AI writing is becoming increasingly formulaic.

• Scott Cunningham on “Memory is the residue of thought” and fixing that when working with AI agents about the problems he is encountering around memory when working with agents and his visual approaches to fixing this.

Longer reads on AI writing that I found interesting/thought-provoking:

• Superslop, and How It Will Kill Your Book Deal is an interesting experiment from Michael O. Church. He says, “I figured it would take about sixty hours and $2,500 worth of Claude tokens to build an AI-generated novel that could break a sophisticated autograder.” However, “The process took five hours and used $209.55 of Claude tokens; its result was a novel that, if presented to traditional publishing with a photogenic, class-fluent white woman as its author, by a literary agent able to get manuscripts read, would secure a book deal with significant support.”

• “Generative AI floods and dilutes the market for books,” by Tuhin Chakrabarty et al., unsurprisingly finds that AI-detected text is increasingly common across the Amazon self-published fiction genres they studied. Books with >25% detected AI text are 20% of titles but account for only 12.1% of sales and 11.3% of revenue, but some sell extremely well, and their share of sales is rising. Meanwhile, the number of titles competing for sales has grown much faster than revenue, and revenue per selling title has fallen. See also: Thad McIlroy’s 6,000-word response.

• Hollis Robbins on The AI campus gender wars ahead: “The most volatile demographic and technological clashes on campus this fall will not be on the streets but in the classroom. Incoming young men, increasingly conservative and highly optimistic about generative AI, will be placed in classrooms with non-tenure-track, overwhelmingly female, politically progressive first-year writing faculty fiercely hostile to AI.”

• Henriette Thiesen, Publishing Director at Gyldendal, on AI and the future of the textbook: “A textbook tells a student: ‘This is what you need to know’. Setting those boundaries is itself a form of expert knowledge, and it is something only people with genuine subject mastery can provide.”

• Dr Sam Illingworth on You Will Be Accused of Using AI. Here Is How to Prove You Wrote It.

Upcoming events:

• Building capacity with AI tools you already have: practical applications for scholarly publishing looks interesting. Sept 16 at 10:30 AM ET.

And finally…

Interesting that the people building language models to automate everyone else’s work still want a human editor for their own content:

End Notes:

• Newsletters might be even more sporadic than usual for a while.

• If you found this useful and feel so inclined, you can always buy me a coffee.

• If you need consulting help navigating any of this, find me at Maverick.

• PLOS’s Redefining Publishing: Practical pathways to open science is long but interesting. “Our transition to Janeway for submissions is one example of this approach. Rather than acting as a passive user of a platform, we are investing engineering effort upstream, contributing code and ideas, and helping to shape the direction of the project alongside a wider community.” I don’t think this is a surprising move, given the arguments in the report or the options available, but I doubt anyone would have considered PLOS using Janeway a serious possibility 10 years ago.

• “Google announced this week that publishers can now remove their content from AI Overviews while staying indexed in regular search. Publishers had demanded this for years. It finally arrived, delivered under pressure from the UK’s Competition and Markets Authority. The toggle went live June 3. And yet it doesn’t feel like relief. Because opting out of AI Overviews means opting out of the surface where search is going. And opting in is the status quo. Content feeds the answers, traffic doesn’t come back. Neither option is a deal.” Using Ezra Eeman’s words here as they sum up the situation nicely.

• The Company of Biologists has published results from its Fast & Fair paid peer review trial. Average time to first decision fell from 37.7 working days to 5.5, invitation acceptance jumped from 23% to 67%, and handling editors found no drop in review quality (Reviewers receive £220 per manuscript). This is a tech-focused newsletter, but including it because human solutions are often easier/faster. [Preprint | Summary in Research Information]

🚀 Start-ups:

• The Dutch Groep Algemene Uitgevers (General Publishers Group) has announced its “Renew the Book 2026” award longlist - seven projects covering everything from AI literary companions to inclusive picture books and cinematic book trailers. Among them is BookStories (openscience.works), which builds open, reusable services to help researchers, libraries, and publishers understand how open scholarly works are being used, shared, and discussed using transparent metrics.
  

📈 Stats:

• The ScholarOne Future of Peer Review 2026 report is out. I’m one of the interviewees, which gives me a vested interest in flagging it, but it does contain some interesting data. For example, reviewer invitations sent by journals via ScholarOne rose from 9M in 2018 to nearly 21.5M in 2025 (some caveats around that stat in the report). Despite working in academic publishing for years and understanding the scale, I still find the size of numbers like this staggering.

• In the first 16 days of April, automated bots generated 5.7 billion hits across BMJ sites, compared with an expected 19.7 million hits in a typical 16-day period (289 times the normal level of traffic). [Ian Mulvany’s Keynote from Force 2026]

• From Springer Nature’s Q1 update: “Our article publication growth of 15% continued to outpace the market, which we estimate grew around 6%.” This is a simplistic analysis, but a 15% growth in output with a 1.7% increase in the implied Research cost base, probably 6-7% once exchange rates are taken into account, and rising profit margins are [quite impressive or troubling, depending on whether you see this as a business scaling well or as more value being extracted from the research community].

• Anthropic research surveyed 1,260 quantitative social scientists on AI use, and the headline found 81% have tried AI chatbots, but only 20% use coding agents like Claude Code that actually write and run analysis autonomously. I think 20% is a high level of usage for a tech that’s really only a year old.

• Ali Ghodsi, Co-Founder & CEO at Databricks: Three years ago, queries coming to Databricks were 100% human. A month and a half ago, it crossed the tipping point past 50%. Today? Over 81% of database queries are executed by autonomous agents, not humans.

• Wiley has acquired Emerald and says it expects about $30m of annual run-rate cost synergies. To put that in context, and if my maths is right, they are expecting annual run-rate cost savings equivalent to roughly $86k per Emerald employee. Emerald has 350 employees and projected FY26 revenue of $85m. $30m divided by 350 employees is $85,714, or about $86k per employee. Wiley is pitching Emerald as part of a larger AI growth model in which trusted research content becomes fuel for data analytics and AI intelligence.

🤖 AI:

• Ian Mulvany and co have written up how the BMJ and AWS Generative AI Innovation Center developed an AI-powered editorial assistant designed to help journal editors screen submitted research manuscripts to make better decisions about which papers to send for further peer review, which to reject, and why.

• Christina Spencer on building AI interfaces at JSTOR

• The Economist is preparing for “a world with two versions of the web” — one optimized for rich, human reading experiences, and another where “agents want clear structure, questions and answers, ideally text,” [Digiday]. This kind of thinking also comes through in Tony Alves’s excellent write-up of the PurePub.ai conference. On the one hand, creating structured content for agents seems obvious to me; on the other hand, we’ve been having this conversation since the semantic web was first mooted. Maybe it will happen this time, or maybe world models will come along, and academic publishers can remain focused on PDF-like containers?

• Mark Riley is thinking along similar lines with news: “For 25 years, most digital news has still been shaped around the article as the atomic unit: headline, image, intro, body copy, related links, comments if you are feeling brave. But the AI interface invites a different behaviour. “Explain this to me.” “What’s the background?” “Why does it matter?” “Who benefits?” That is not search. It is not social. It is not even really aggregation. It is news as an explainer engine”

• The Hand That Feeds: Publishing’s AI Hypocrisy Problem by Mark Williams. “Publishers have constructed, with meticulous care, a moral circle around creative labour. Inside the circle: authors, illustrators, narrators, translators - the human artists whose displacement by AI is framed as an existential crisis, a theft, an assault on imagination itself. Outside the circle: engineers, warehouse operatives, phone operators, metadata specialists, marketing copywriters, junior editorial assistants - the human workers whose displacement by the same AI is framed as efficiency, innovation, or simply not mentioned at all.”

• Simon Linacre from Cabells on The end of the papermill? “One consequence of this increased use of AI to produce articles is that usage of paper mills may actually decline. The reasoning is that, assuming some authors follow a path of least resistance in their publishing strategy, they will see that the benefits of paper mills (ready-made articles, supplied for a price) are less than the benefits of using AI to write something for them (ready-made articles, supplied for free).”

• Thibault Geoui on A system of trust for science: rethinking validation (/review) in scientific publishing (Via Chris’s must-read Scalene newsletter)

• I agree with much of what Todd Toler says in his Scholarly Kitchen interview, particularly the comments about proprietary solutions leading to fragmentation and silos. I don’t entirely agree with this: “AI agents struggle to access PDFs and extract information from figures and tables”. I think the latest models are getting good at this, but they aren’t perfect yet.

• A Snapshot of GenAI Tools for Research | HKUST Library is a really good summary, highlighted by Aaron Tay in AI academic search needs better frameworks for understanding and evaluation. These three librarian projects are a start

• 🌎 Ithaka S+R has produced a LibGuide focused on the Environmental Impacts of AI

📅 Events:

• Cambridge AI meetup, 25 June (webinar): AI Hallucination - why it matters and what can be done. Speakers: Damien Charlotin on his database of legal hallucinations, Richard Sloggett (Referonix) on catching AI errors in sales proposals, and Nick Morley (GroundedAI) on how well frontier LLMs actually avoid hallucination.

• BookTech Remote Con, the first online event of BookTech Collective, goes live. “A curated series of keynotes, panel discussions, and strategic conversations exploring the future of publishing in the age of AI. Featuring prominent voices from across the global publishing industry.”

📚 Longer reads:

• Tim Ferriss (The 4-Hour Workweek), on Has AI Already Killed How-To Nonfiction? Sales Trends, My Personal Data, and What It Might Mean for the Future

• Building Search for AI Agents with Exa CEO Will Bryk. “Bryk shares his views on AI-native products, the future of information discovery, and why some of the most important problems in technology can ultimately be framed as search problems.”

• 🎧 AI Eats the World? A Reality Check with Benedict Evans. “Erik Torenberg speaks with tech analyst Benedict Evans about the current state of AI, what has changed over the past year, and which questions remain unanswered. The conversation covers coding agents, foundation models, AI infrastructure spending, software economics, and the tension between today’s AI excitement and the long-term realities of technology adoption.” Recommending because it’s so grounded/cuts the hype about what AI can currently do.

And finally…

Alberto Romero (@thealgorithmicbridge): "Museum of Meaningless Metrics"

End Notes

• If you found this useful, you can always buy me a coffee.

• If you need consulting help navigating any of this, find me at Maverick.

You can catch me later today at PurePub.ai, where Ian Mulvany, Andrew Smeall, and Ann Michael will be exploring the emerging opportunities around AI in What should publishers be talking about?

🆕 News:

• The PurePub.ai conference is up and running this week, and there’s still time to catch a couple of the sessions live.

• STM is seeking feedback on its new brief Toward Responsible Use of Research Content in Generative AI, looking at how scholarly values and standards could be respected when using research content in and by GenAI systems and applications.

• arXiv’s policy on the use of generative AI is, I think, a very sensible approach: “Our Code of Conduct states that by signing your name as an author of a paper, each author takes full responsibility for all its contents, irrespective of how the contents were generated.” Reading through the comments, especially from some of the more indignant researchers, is a treat. What, I have to check the paper that bears my name?

• Scopus AI is now called ‘AI Discovery’, a name that must have taken the marketing team hours of hard thought to come up with 😉.

🚀 Launches:

• Google’s Deep Research Max announcement explicitly names open-access peer-reviewed journals as one example of the authoritative sources its reports can draw on. I don’t keep a close eye on this, but is this one of the first mainstream AI product announcements to foreground open-access peer-reviewed content as the content pool?

• Sci-Hub (widely treated by courts and publishers as a copyright-infringing shadow library) has launched a new AI-powered research assistant called Sci-Bot (sci-bot.ru). Mixed reviews from scientists on Chemical & Engineering News, who note that Sci-Hub’s increasingly limited access to paywalled literature reduces the value of the tool. Nick Morley from Grounded AI analysed the references and found that of the 514 references created, nearly 30% had some form of author misattribution.
  My own quick test for a history of Digital Science was mixed; it covered the key points, nothing beyond 2023, and included some absolute clangers. For example, Daniel Hook wasn't the founding CEO. Around 50% of core claims and references were accurate, 30% inaccurate or misleading, 20% would need more careful checking.

• Springer Nature’s Cureus is launching the Journal of AI-Augmented Research in May 2026 as “an open and trusted home for researchers using AI as part of their research practice, with clear standards for transparency, reproducibility, and human-verified validation”. It was only a matter of time before research created using AI, and its associated OA fees, found new homes at major publishers.

• ReviewBench is an open-source, venue-agnostic framework that compares human and AI reviews across structure, alignment with a paper’s major claims, impact, and critique category. [Preprint | GitHub]

🖧 Fabricated citations:

• My feeds are full of reports about fabricated citations in research papers. I am quite curious to know how this October 2025 (!) Frontiers paper got published. 🤷

• Max Topaz’s CITADEL site is a great piece of data science and visualisation summarising their research findings about fabricated citations.

• Retraction Watch has a good summary of Topaz et al.’s research. I thought this was a good comment: “For instance, citations that play a minor role (e.g., placing the reported study within the context of existing literature; serving a rhetorical function) and which can be withdrawn without affecting the argument made or the validity of the results can be addressed in a correction that is transparent with respect to the authors’ lapse in vigilance.” (It’s surprisingly easy to get a fabricated reference when you do something like ask for the landmark study in JAMA about X.)

• Tim Elfenbein asks what I think is the more important question: “Machine fabricated citations are a problem, but they are an extension of lots of other all-too-human problems w/ citations that have gotten little attention. What is worse: a completely fabricated citation or a misrepresentation of an existing cited work?”

• A Nature analysis estimates that more than 110,000 scholarly publications from 2025 alone may contain fabricated references generated by AI. [Summary on LinkedIn]

• I am feeling slightly wistful. Back in the day, there were teams of people (like me) working in ‘Electronic production’ departments manually checking that every URL in a reference list resolved to the right paper before it went online. Utterly mind-numbing but quite relaxing work that has largely been automated away (thankfully!). It is hard to believe the work that went into getting papers online 20+ years ago, the levels of QA done because of the fear that something might be incorrect online (which, to be fair, it often was).

🤖 AI:

• IOPP’s new Duplicate Review Checker uses machine learning to identify reviewer reports with significant content overlap. Since its 2024 pilot, it has processed around 500,000 reports going back to 2020, flagging matches above a 60% similarity threshold. It catches both reports reused across multiple manuscripts and reports submitted under different reviewer names, with flagged reports routed to the Research Integrity team for review. Around 2,500 duplicates have been caught so far.

• 🎙️ Prathik Roy, Product Director for Data and AI Solutions at Springer Nature, on The Product Experience podcast, covers a lot of ground, including: “You’re effectively at an inflection point where you need to start thinking about whether you are building for humans to consume content via machines instead of humans directly consuming content off of your platforms.” So, how do you quantify the value of content when nobody visits your site anymore? Prathik walks through tracking content into AI pipelines, measuring what proportion of an LLM’s answer came from your sources, and eventually pricing by outcomes rather than access.

• 📊 COUNTER has released its Best Practice on Generative and Agentic AI usage metrics. See also the Research Information article by Tasha Mellins-Cohen.

• The Foresight Institute is funding projects on this premise: “In addition to applying AI to specific problems, we need better platforms, tools and data infrastructure to accelerate AI-guided scientific progress generally. Similarly, to get our sense-making ready for rapid change, we are interested in funding work that applies AI to improve forecasting and general epistemic preparedness.”

• 🗎 Talip Gönülal, Ramazan Güçlü and Salih Güçlü explore AI In Academic Publishing for Non-Native English Speakers: The Good, the Bot, and the Ugly: “… the good, reducing linguistic inequalities by improving paper quality and decreasing language-related challenges; the bad, involving inaccurate or misleading AI suggestions, over-reliance on AI tools, and diminished engagement with manuscripts; and the ugly, characterized by failure to disclose AI use, lack of clear guidelines for responsible AI integration in research, homogenization of academic writing, and the emergence of new forms of inequality.”

• James O’Sullivan on The AI detection delusion: “...the solution to one form of technological recklessness cannot be another. Detection tools give their users the feeling of objective certainty while delivering probabilistic guesses, and in doing so, they can cause real harm to real people—students who lose marks or face disciplinary action, professionals whose reputations are damaged, and writers whose command of English is held against them by an algorithm that mistakes simplicity for artificiality.”

• Dave Flanagan at Wiley makes a related point in Provenance, not detection, arguing that the right question for research integrity isn’t “did AI write this?”, it’s “can we trust this author?”. Note Chris Reid’s comment: “Trust will be at the author layer.”

• Dustin Smith in Claude Code Is a Rogue Wave for Publishers writes that AI chatbots have driven a ~30% increase in manuscript submissions, but that AI agents will make that look small. If 15% of researchers adopt agentic practices by 2030, submission volumes double. The third-order consequence will be that the most valuable capability shifts from dissemination to rejection, routing, and authentication.

📚 Longer reads:

• 🎙️ I enjoyed and learnt a lot listening to this podcast with Andri Johnston, Digital Sustainability Lead at Cambridge University Press, talking in part about the environmental impact of AI.

• Nova Techset CEO Yogesh Agarwal says we should pay more attention to the unglamorous operational work holding scholarly publishing together: editorial coordination, workflow handoffs, quality checks, all the stuff that doesn’t show up in dashboards but quietly keeps the system functioning. As submission volumes grow and AI gets layered in, Yogesh argues that this invisible work is multiplying rather than going away.

• 🗎 Gary Marcus flagged a new Nature paper at the TAIS meeting last week. The paper, “State media control influences large language models,” found that human annotators rated Chinese-language answers as more favourable to China in 75.3% of comparisons. This paper is about politics, but I think the findings might apply more broadly, for example, in fields where the Chinese-language and English-language framings of a topic might diverge, such as Traditional Chinese medicine. If you ask about acupuncture or moxibustion in English, and then in Chinese, would the same model give you different answers? Not because the underlying evidence differs but because the framing in Chinese-language sources tends to be more positive.

• 🔎︎ Aaron Tay on “We’re Good at Search”… Just Not the Kind That the AI era Demands — a Provocation, on how AI is changing librarians’ roles.

• 404 Media reports on Arizona State University’s rollout of “a platform called Atomic that creates AI-generated modules based on lectures taken from ASU faculty by cutting long videos down to very short clips then generating text and sections based on those clips.” I can see why this looked good on paper, and I can also see that this would be phenomenally difficult to do well with the tech we currently have.

• Matthew Scott Goldstein on the scraper economy: “We track 15 companies systematically extracting publisher content to power AI systems. They are not all doing the same thing — but the result is identical. Publisher content flows in. Publisher compensation flows out. Here is how each tier works and what it looks like in practice. The market these 15 companies serve hit $1 billion in 2025 and is projected to reach $2 billion by 2030. Not one dollar of that flows back to publishers.”

And finally…

📖 Apparently there are enough novels about research integrity to fill a literature review. Lex Bouter has read 35 of them from the last 100 years, "with special attention to the research integrity topics, its drivers, and its consequences". A summer holiday reading list? ⛱️

End Notes

• If you found this useful, you can always buy me a coffee.

• If you need consulting help navigating any of this, find me at Maverick.

🆕 News:

• Kudos is inviting publishers to join/sponsor its “Taming the Crocodile” study, which examines the growing impact of zero-click search and AI summaries. The research aims to help publishers protect usage, revenues, and research integrity in an increasingly click-free world.

• Rob Johnson reports that Elsevier is shutting down Researchfish (the platform tracking research outcomes for UKRI, NIHR, and 90+ other funders). I’d completely forgotten about “Researchfishgate” when they responded to critics on Twitter with a vaguely threatening message: “We understand that you’re not keen on reporting on your funding through Researchfish but this seems quite harsh and inappropriate. We have shared our concerns with your funder.”

• The UKRI’s Metascience Unit has announced “a new funding opportunity: a Scientometrics Sandpit. We are convening researchers and practitioners to co-create projects that develop, validate and critically assess novel scientometric indicators for use in research assessment.”

• What does $200,000 buy you in the AI training data market today? Apparently, high-speed access to a pirated library of more than 60 million books, payable in crypto.

• GetFTR now lets AI-powered research tools check users’ access rights in real time, using the Model Context Protocol to route them to the appropriate full text.

• Retractions go mainstream. The Sunday Times [Paywall] reports that 500 papers are now being retracted per month. The article covers overwhelmed peer review, AI-generated slop, paper mills, editor bribery, and incentive systems rewarding flashy findings.

• 67 Bricks’ report on B2B publishing tech maturity shows AI adoption is high, but impact is low, arguing that competitive advantage will only come from connecting data, treating AI as a real transformation, and shifting from static publishing to signal-driven, workflow-embedded products.

🚀 Launches:

• OpenNote has launched Papers, an AI-native academic writing suite built on Feynman, Opennote’s teaching-focused AI that helps students learn by reasoning through problems, using visuals, and adapting to their progress.

• OpenAI launched Prism. Is it an Overleaf killer? I suspect not. Overleaf has tens of millions of users, so it’s unlikely they’ll all migrate to a new product unless it does something genuinely new. I see this less as a bid to replace Overleaf and more as an attempt to own the scientific workflow, or possibly the scientific process, rather than to acquire training data. [Also see discussion on LinkedIn and note on Bohrium, another AI Science solution, below]
  
  I’m more struck by the contrast between these two videos and what they reveal about how science and writing actually happen. In the first, Kevin Weil delivers a full-on Silicon Valley, fast-talking GTM Sales pitch. Fast, confident, visionary. AI will accelerate science just as it has accelerated coding, complete with big comparisons and future visions of robotic labs. In the second video, a quiet, slow-paced demo of Crixet (the previous name for Prism), you see the reality of human work. Victor Powell, the developer, is in his kitchen, minding the baby (the really high-value work), while talking to an AI tool to handle writing a scientific paper. As a non-LaTeX person, I’ve always found it arcane and faintly archaic. I’m not convinced anyone will need to write in LaTeX in ten years’ time, but then again, many arcane systems survive far longer than anyone expects.

📈 Interesting Stats:

Including simply because the numbers caught my eye:

• Between 2020 and 2025, submissions to NeurIPS increased by more than 220% from 9,467 to 21,575.

• Cabells now includes over 20,000 journals in its Predatory Reports database. 😯

• DOAJ on scraper bots: “The last six months of 2025 saw a 419% increase over the same period in 2024, culminating in a single day in mid-November where our traffic spiked to 968% greater than a year earlier.”

🔍 Search:

• 79% of almost 100 top news websites in the UK and the US are blocking at least one crawler used for AI training. Reports of SEO’s death are greatly exaggerated, though traffic to reference materials is down by 11% 📉

• “Large [news] publishers who blocked AI bots from crawling their sites saw total website traffic drop by 23% — and that’s not just from removing bot visits. Human traffic declined by 14% too.” [This is a summary generated by Hacks & Hackers AI Papers Explained, their experiment in using AI to help translate the latest AI research from arXiv into plain language for journalists and technologists, original paper]

• Google says it’s exploring new controls that would let websites specifically opt out of having their content used in Search’s generative AI features like AI Overviews and AI Mode, as a response to UK regulatory pressure, but details and timing are still unclear.

🤖 AI:

• John Fairbanks from Silverchair, on detecting GenAI-generated content, “Last week, I watched a room full of publishing professionals run the same abstract through GPTZero, CopyLeaks, and Pangram. One tool said 100% human. Another said 100% AI. Same text.”

• Darrell Gunter argues that scholarly publishing needs a neutral governance body for the AI age. I am not convinced we need this; my thoughts here.

• Microsoft has announced its Publisher Content Marketplace (PCM), a solution intended to give publishers a new revenue stream by providing AI systems with scaled access to premium content. Ezra Eeman captures the current landscape in this infographic:

  

  Not mentioned in the infographic is Cashmere. Cashmere founder Jonathan Woahn has 2 articles in the Scholarly Kitchen covering why AI Isn’t Going to Pay for Content … At Least Not How You’re Hoping It Will, and The Path Forward.

• Geoffrey Bilder’s Instead suggests that websites should provide AI bots with a direct link to a downloadable, static version of their content, reducing the need for repeated crawling of live pages.

• Moara.io compares Claude against Elicit, Consensus, SciSpace, Undermind, and Scholar Labs. Claude came out on top, ranking higher than specialised platforms like Elicit and Consensus in an evaluation of the top 10 search results on the economic impacts of climate change. [Also see commentary by Aaron Tay & others on Bluesky]

• Wikipedia has agreed licensing deals with several major artificial intelligence companies. The Wikimedia Foundation will now be compensated for access to its data, which will be used to train AI systems. TNPS on lessons for book publishers

• Haseeb Md. Irfanullah argues that 100% AI-Reviewed Preprints are the Future of Open Research. I agree that preprints are increasingly accepted and unavoidable in some fields, and that human-run journals and peer review have problems, but I don’t agree that the current system is inherently unjust. I agree that AI-based review can plausibly operate at scale, but is this desirable? Journal publishing, and preprints to a lesser extent, is relatively decentralised, with evaluation distributed across multiple communities and institutions. Large-scale AI review would tend toward centralised criteria, infrastructure, and decision logic, creating risks of widespread error and reduced transparency. Replacing human peer review would relocate decision-making power to AI systems, raising questions about who would own, govern, and be accountable for these systems. [See also comments from William Gunn about peer review at the bottom of this article on AI Disclosures]

• SSP’s first Pulse Check poll covers “AI in Scholarly Publishing”. Conclusion: “The prevailing mindset is neither resistance nor blind enthusiasm, but cautious experimentation shaped by deep concern for scholarly values”:

• A personalisation experiment from Amy Webb that really does deserve a moment of your time. “We are on the cusp of aggressive algorithmic personalization, and I think it’s important that we stop and ask how and why decisions are being made. To me, this feels like an early signal of a larger shift: audience-aware intelligence that adapts framing as much as facts. That has implications for leadership communication, media, and trust -- especially as these systems move from assistance to influence at scale.”

• Bohrium is one of many companies building infrastructure for agentic science at Scale [LinkedIn note | Paper].

  “Bohrium—a platform that turns scientific assets—papers, code, instruments—into things agents can call, observe, and learn from. Think of it as Unix for experiments. It understands not just files, but the semantics of data; it supports closed loops between models, simulations, and lab benches; it lets different disciplines share a common workspace.
  Above this sits SciMaster—an orchestrator that coordinates long-horizon research across reading, computing, and wet-lab work. SciMaster doesn’t replace scientists. It provides the scaffolding within which scientific work can be executed, inspected, validated, and improved.”

📚 Longer Reads

• 🎥 This short docu-film sets the stage for Frontiers Science House at Davos. It tells a compelling story about openness, framed from the perspective of a publisher that materially benefits from how openness is currently implemented. Note, the shift in focus to Fair Squared, a data-management platform positioned as making scientific data “certified” and “citable,” promising researchers a new way to build careers while enabling data aggregation in the name of faster problem-solving.

• Anil Dash on How Markdown took over the world.

• The Age of Academic Slop is Upon Us - by Seva Gunitsky includes the wonderful made-up definition: “Automatenwissenschaft (n): A flood of AI-generated, technically adequate papers that nobody reads.”

• 📄 Another lovely term, PISS: “One person PISSing in the pool of science? Negligible. Tens of thousands? The water takes on a different hue...” Hanson et al. analysed over 100,000 special issues and found that 1 in 7 guest editors publish in their own issues — a practice they dub PISS (Published In Support of Self). The result? Up to 43,000 excess articles.

• 🎧 Michael Nakan CEO of Comixit! is using AI to transform the legendary Beano into “frictionless, digital-native webtoons”

• Jane Hammons in “Outdated Notion? Teaching Plagiarism as Theft,” challenges the metaphor of plagiarism as “theft”, arguing that this framing is legally inaccurate, culturally narrow, and pedagogically harmful and proposes a shift toward more supportive language that treats citation as an academic practice rooted in respect and collaboration, rather than criminality.

• 📄 A study of ~280,000 academic YouTube transcripts found that the frequency of certain words, such as delve and adept, shows a statistically significant upward trend starting after the release of ChatGPT in November 2022, with estimated increases of roughly 35%–51% over the ~18 months of post-release data.

• 🎧 Andrew Smeall: using technology to effect business change | Thinking about Digital Publishing

• Steve Smith argues that the traditional PDF is becoming obsolete as the core scholarly unit of record because it obscures machine-readable structure and context, and that publishing needs to shift toward “Knowledge Objects” that encode claims, context, and provenance explicitly for AI to reason reliably with research information. But how do we actually get from today’s messy, incentive-driven publishing workflows to interoperable, trustworthy “knowledge objects” at scale?

• Rick Anderson revisits his 2014 “quiet culture war” thesis to argue that divisions within academic libraries over institutional service versus global open access have intensified. He argues that the open access movement has grown more ideological, with little room for dissent, and that professional pressures stifle honest debate. (I have a feeling this article might resonate more with publishers than librarians. For all the good that OA provides, I think many publishers look back with a sense of wistfulness about the world that was.)

• Mark Hahnel argues that the next phase of open data must focus on quality, incentives, and stewardship, supported by emerging technologies, if it is to deliver meaningful value to research and society.

• Azeem Azhar and Hannah Petrovic on How “95%” [of organizations are getting 0 return from AI] escaped into the world – and why so many believed it. The tale of where this stat came from is the interesting part of the read.

• Anita Coleman’s Infophilia post explores how library AI voices show up on Substack, examining how information behaviours and platform dynamics shape visibility in AI literacy discourse.

And finally…

• Andrew Maynard asks, Can modern scholarship escape AI? in a new (sadly unpublished) paper and provides a succinct answer:

End Notes

• If you found this useful, you can always buy me a coffee.

• If you need consulting help navigating any of this, find me at Maverick.

• Looking for my AI-powered use cases for Academic Publishing? Find the list here

🆕 News

• 📅 The Research Data Alliance is relaunching its Research Data Policy Interest Group as a more global forum for discussing research data policies. More information and a registration link for the webinar on January 20, 2026 11.00 GMT / 12.00 CET.

• 💰Sage Publishing has sold Lean Library, a digital tool aiding access to library services, to its management team, including longtime director Daniel Horvath and tech lead Nenad Vidic.

• 🤖 What’s more of a problem for scholarly publishing - bad bots or bad humans? Results from the STM Research Integrity Day panel.

• 📊 More than half of researchers now use AI in peer review workflows, according to new data from Frontiers

• 🔍Crossref explains how to spot a fake DOI, for example, DOIs cannot begin with anything other than 10, cannot have three-digit prefixes, cannot start with 0, and always start with a 10 followed by a period.

• 🗳️Deanta’s excellent Trends in Academic Publishing, 2026 is up and running.

• 📈The OA Switchboard’s 2025 report shows it sent over 2.3 million messages between 42,000+ connected organisations, including 42 publishers, 21 consortia, and 280 institutions and funders. (The OA Switchboard is a neutral, non-profit platform for sharing open access publication metadata, founded in October 2020.)

🚀 Launches

• CABI has launched SDGenie (beta), a prototype tool that matches research abstracts to specific SDG targets, helping researchers show how their work supports global goals. It’s free to use. Give it a go and share your feedback via the short survey. [Late to this - it launched in November]

• WyderNet is a new AI-powered platform that embeds librarian expertise into students’ digital workflows by interpreting course materials in context. It maps assignments and syllabi to authoritative library resources, ensuring credibility wins where convenience once ruled.

• Andrew Ng has launched Agentic Reviewer to give feedback on research papers (works best on AI papers). In testing, its comments aligned with human reviewers slightly better than two humans did with each other (Spearman 0.42 vs 0.41), but critics argue this reflects mimicry, not judgement, because the tool copies the tone of past reviews.

• Trilogy has launched Manuscript AI, a new tool that helps book publishers, commissioning editors, and literary agents sort through unsolicited manuscripts and the slush pile. It rates submissions on sales potential, likely reader reviews, genre match, and writing style, using data from 2.7 million copyright‑cleared and public domain books. The tool covers major categories, including memoir, nonfiction, drama, romance, and science fiction, and aims to spot any hidden gems.

• TIB AIssistant is a new open-source platform built using Next.js and GPT-4o mini, that stitches together a library of AI assistants for every stage of the research process. It integrates external scholarly tools like ORCID, Crossref, and Semantic Scholar, and lets users export machine-readable provenance data with RO-Crate. [Article/ArXiv preprint]

🤖AI

• 📅 Impelsys is hosting a panel discussion on Publishers’ Opportunity or Dilemma? Licensing Content for AI Training February 05, 2026 10:00 am EST | 4:00 pm CET

• Kent Anderson asks, Can you retract an article from an LLM? Kent Anderson explores NEJM’s AI Companion and finds that it misrepresents scientific findings, confuses authorship, and fails to flag retracted papers. Kent argues that opaque LLM technology, poorly understood even by the publishers adopting it, is not fit for purpose and risks eroding trust in the scientific record. See also David Worlock’s response and my response.

• 🔍 Aaron Tay’s exploration of a very modern user crisis: what are we actually meant to type into AI search boxes?

• Peter Schoppert summarises Roy Kaufman’s argument on whether training LLMs on open-access content is legally sound, particularly when attribution is removed. He also revisits Google’s 2018 Talk to Books experiment, an early example of using an LLM as a conversational search tool that predates ChatGPT by several years.

• Dr Stylianos explores the flawed promise of AI writing detectors, which rely on statistical patterns like “perplexity” (how predictable or uniform the word choices are) and “burstiness” (variation in sentence length and structure) to flag machine-generated text. He argues that as models evolve, their ability to mimic human nuance increases, which makes AI detection tools less reliable and reduces their value. [I use Pangram’s service fairly regularly and like it, but I’m aware that when I submit entirely human-written work about AI, it’s often flagged as 100 percent AI, with many common terms marked as suspicious.]

• UKRI is testing whether generative AI can predict reviewer scores on 1,000–2,000 past grant proposals, without seeing the outcomes. Led by Mike Thelwall, the project explores if AI could support peer review by easing admin, breaking ties, or acting as an additional reviewer.

• The FT argues that tools like Claude Code are reshaping academic research by automating once-specialised tasks. While this boosts reproducibility and lowers production costs, it may encourage more marginal research activity with greater output but not necessarily greater insight. As technical barriers fall, will intellectual originality become the key differentiator?

• Mark Hahnel argues that research data should prioritise machines over humans, enabling AI systems to accelerate discovery. That means ditching PDFs and vague metadata in favour of structured, queryable, computational assets.

• Problems with Publishers Moving to AI-Based Production

• Springer Nature is pulling nearly 40 papers after researchers used a flawed dataset of children’s images from autism-related websites without proper consent. [There are a lot of dodgy datasets out there (e.g., DukeMTMC4ReID, 80 Million Tiny Images, Brainwash, etc), and publishers probably need to get better at identifying research that has used these.]

📚 Longer reads

• Tim O’Reilly’s AI and the Next Economy is a clear-eyed look at the economic risks of AI, framed through a Blakean warning. “We may be building the engines of extraordinary productivity,” he writes, “but we are not yet building the social machinery that will make that productivity broadly usable and broadly beneficial. We are just hoping that they somehow evolve.”

• This Euractiv article tackles how AI is rapidly transforming journalism, faster than most institutions or the public realise, and why trust is becoming the central issue, not just disinformation.

• A new preprint by Zhao & Berman finds that LLMs are already disrupting the online news ecosystem. Since August 2024, news publishers have seen a 13% drop in traffic, likely due to AI tools summarising content without linking back. Sites that blocked AI bots like GPTBot saw even steeper drops, including a 14% drop in human visits and a 23% decline in total traffic. Attempts to block AI crawlers appear to reduce both total and human traffic further.

• Ask A Librarian offers a critique of how generative AI reshapes student research by replacing the exploratory process of scholarly search with answers that feel perfect but may be fabricated. It draws on real librarian experience and Gen Z search habits to argue that learning happens not through instant results but through questioning, frustration, and discovery.

• Professor and economist Joshua Gans spent 2025 “vibe researching” (using large language models to quickly draft, revise, and submit academic papers). While the process boosted output, he found it also encouraged bloat, pushed weaker ideas too far, and made it easy to mistake fluency for substance, with confident-sounding but shaky results often slipping through until peer review.

• In The Lineages and Inheritances of Shadow Libraries and their Documentation, Martin Paul Eve considers the fractured legacy of shadow libraries after the takedown of Library Genesis, focusing on how successor projects like Anna’s Archive try to stitch together and tidy up metadata from scattered sources.

• Rosalyn Metz questions the uncritical worship of “openness” in library and digital infrastructure communities and suggests we rethink openness not as a product to give away, but as a shared responsibility requiring consent, stewardship, and structural care.

• Eleonora Dagiene describes how efforts to fix a broken publishing system in Lithuania ended up weakening it further. Reform policies aimed at raising standards instead undermined local journals, leaving researchers with little choice but to rely on commercial platforms. What started as a response to poor practices became a deeper loss of trust and control.

• Kevin Baker’s “Context Widows” essay argues that questions about whether LLMs can “do science” miss the point; what matters is how these tools are integrated into an institutional landscape dominated by citation metrics and other evaluative measures that have long distorted scholarly priorities. Rather than challenging this system, large language models function as amplifiers of existing incentives, rewarding volume and form over substantive contribution.

• 🎥 Oxford academic Professor Patricia Kingori explores Kenya’s shady, multi-billion-dollar fake-essay industry, where educated Kenyans secretly write papers for students around the world. The Shadow Scholars is ponderous and slow, but tackles a complex issue.

And finally:

🤦‍♀️ In a new low for academic integrity, Springer Nature’s £125 AI ethics book is riddled with fake citations, including references to journals that do not exist. The publisher says it is “investigating”.

End Notes:

• 💷 If you found this useful, you can always buy me a coffee.

• 🔗Looking for my AI-powered use cases for Academic Publishing? Find the list here.

• 🤖If you’re interested in AI, you can subscribe for free to my other Substack, GenAI for Curious People, to see entries for my new book.

• 🆘If you need consulting help navigating publishing technology or AI, find me at Maverick.

🆕News

• 🦹 A bug in OpenReview allowed author, submissions, reviewer details, and their reviews to be leaked, and a dataset of 10,000 ICLR submissions circulated. The Glass Wall Shatters: A Professor’s Reflection on the ICLR 2026 Breach is a good summary of the implications and the fragility of the peer review system. I’ve only just twigged that this is where the data for Pangram Labs’ “21% of the 75,800 peer reviews for ICLR 2026 were fully generated by AI” came from.

• 📈 This stat from Richard Wynne caught my eye “...during the next three years scholarly journals will send out an additional 150 million invitations to volunteer peer reviewers OVER AND ABOVE the current baseline”. I think you can quibble with the calculations behind this stat, but it’s still a thought-provoking number.

• 🙄 Springer Nature flags paper with fabricated reference to article (not) written by Retraction Watch’s cofounder Ivan Oransky. Retraction Watch says “The paper with the nonexistent reference, published November 13 in DARU Journal of Pharmaceutical Sciences, criticizes platforms for post-publication peer review — and PubPeer specifically — as being vulnerable to “misuse” and “hyper-skepticism.” Five of the paper’s 17 references do not appear to exist, three others have incorrect DOIs or links, and one has been retracted. ”

• 📅 UX in Publishing is a new cross-industry community of practice for anyone interested in user experience in academic publishing. Next meeting 8th December, 2:30 pm GMT (online) on Demonstrating impact and Linking UX to business goals.
  

🚀 Launches

• VeriMe is a new global digital identity verification service for researchers run as a limited cooperative association based in Colorado, US. You can take their survey to help them better understand experiences with identity management and privacy.

• openRxiv announced is piloting a reviewing tool from q.e.d Science to generate rapid AI-generated feedback on biomedical manuscripts. See also: AI reviewers are here — we are not ready

• Elsevier has launched LeapSpace, the “...integrated workspace will combine the broadest collection of trusted scientific content with responsible AI to help researchers move from curiosity to discovery faster” [Press release].

• Google Scholar has launched Scholar Labs, an AI-powered Scholar search that is designed to help you answer detailed research questions. [Google’s blog post][Aaron Tay’s review]

• A British start-up, Books By People, is introducing an ‘Organic Literature’ certification to help readers identify books written by humans rather than machines.

• 🎧The Informed Frontier. A new podcast “examining innovation at the intersection of science, technology, and media. Season 1 explores AI as a ‘reader’ of content, and how this fundamental shift is changing the way we access, interpret, and share information. Hosted by Jessica Miles, each episode looks at the philosophical considerations and practical consequences of this disruption, focusing on scientific and scholarly content.”

• 🎧 Deanta’s Trends in Academic Publishing podcast features Patrick Shafe, talking to Jennifer Regala (Wolters Kluwer Health), Tim Williams (Edward Elgar Publishing), Nicola Ramsey (Edinburgh University Press), Gillian Shasby (Journal of Neurosurgery) and Darren Ryan (Deanta).

• Scholarly Futures is a new weekly Substack from seasoned publishing voices, offering shared pondering on AI, integrity, policy shifts, and the unknowable roads ahead.

🤖 AI

• Ann Michael‍ from AIP Publishing on What We Learned from the Peer Review Assistant Trial. “Citation checking, metadata validation, and ethics compliance stood out as areas where automation added genuine value. … Not every feature landed perfectly. The manuscript digest didn’t always capture a paper’s nuance, and there were some latency issues - the full report sometimes took longer than ideal to generate. Addressing the latency issue is a top priority to allow a more effective fit into editorial workflows.”

• 📈 Nick Morely, from Veracity, asks: What is the cost of citation checking? “Est 75s (1.25 mins) average per ref (or 41 clicks) x 30 refs = 37.5 mins per paper (1230 clicks)”.

• According to Amazon, “less than 5% of titles” on its site are available in multiple languages, so it’s launched Kindle Translate.

• “Gross, disrespectful, and infuriating! The Association for the Advancement of Artificial Intelligence (AAAI)’s move to include “AI”-produced reviews into their selection process is UNACCEPTABLE.” says Dagmar Monett on the introduction of AI reviews. You can’t always please everyone, but you can, like AAAI, be transparent about the process and how AI is being used, as the FT article below explores.

• I think Tricky Trade-Offs on a Transparency Spectrum: How the Financial Times Approaches Transparency about AI Use in News is a helpful read for anyone thinking about policies around the use of AI and what to consider.

• Notes from Nancy Roberts about Deanta’s roundtable about The Future of AI and Its Implications for Academic Publishing. I think this point does need a lot more thought: “Google Zero – the idea that Google’s AI summaries will soon eliminate the need for anyone to click through to a website – is a concern for any of us who publish high value, peer reviewed content. There are no easy answers here but it seems to me that we need to think about the implications of this as an industry, and in conjunction with the wider ecosystem; how do we measure impact of academic research without the traditional metrics of downloads and citations? What are the implications for our publishing platforms? For academic libraries? For our authors?”

• 📈 Steve Cramer in This Liaison Life highlights this stat and beautifully diplomatic reporting language from the UX Team in Design and Discovery at the University of Michigan Libraries survey: “Undergraduate students led overall AI usage at 72.1%, graduate students followed at 62%, while faculty trailed at just 39.4%. Librarian participants reported the lowest usage of AI at just 34%. This inverse technology adoption pattern means librarians have opportunities to deeply understand student AI usage and to integrate much needed guidance for more ethical and effective use of AI tools within library instruction and consultations.”

• The Journalism Benchmark Cookbook is creating a community-oriented AI benchmark for journalism. Is anyone creating community benchmarks for LLMs in journal publishing? Seems like a good idea?
  

🔎 Search

• I enjoyed Michael Upshall’s summary of the Search Solutions Conference, especially the closing point: ““Is AI making search better?”. One respondent put it nicely: “Search has fundamentally changed because of AI, but I am confused about what search is now”. ”

• “We’re Good at Search”… Just Not the Kind That the AI era Demands - a Provocation. I think this is a must-read about the uncomfortable reality that traditional search expertise doesn’t help you evaluate AI-powered tools - and librarians are looking at the wrong metrics entirely.

• Andrew White on how The Edison API can be used to convert PowerPoint presentations to scientific claims and then diligence each claim to see if they are contradicted in literature/patents/clinical trials is a nice demo of how scientific agents can be used.

📚 Longer reads

• Short but challenging read about trust and the internet by Jimmy Wales [Behind a paywall]: “The challenge of our time is not that information is scarce but that authenticity is. Important aspects of the early internet succeeded because people could trace what they read to another human being, even if the other human being was operating behind a pseudonym.
  The new internet must restore that chain of custody. We are entering an era when machines can mimic any voice and invent any image. If we want truth to survive that onslaught, we must embed transparency, independence and empathy into the digital architecture itself. The early days of the web showed it could be done. The question is whether we still have the will to do it again.“ 🎧Or listen to HBR’s interview with Jimmy Wales on the same topic.

• Book Publishing & Technology: The ONIX 3 Debacle by Thad McIlroy, “Will the management of publishing companies figure out ways to harness the power of AI, while mitigating its drawbacks? There are some positive early indicators. But publishing’s historic discomfort with technology seems likely to block the industry from robustly integrating AI into current workflows. The implications are concerning.”

• 🎧TWiST podcast interviews Medium CEO Tony Stubblebine on the company’s new Really Simple Licensing (or RSL) initiative, and how it’s helping to compensate writers for their content and Human Native CEO Dr. James Smith on AI licensing, why a writer’s content isn’t just DATA to them, and why the company is pivoting away from its old marketplace model.

• Deepak Varuvel Dennison, a PhD student at Cornell University, explores the knowledge that is missing from LLMs and the complexity of including that knowledge: Generative AI has access to a small slice of human knowledge. “The marginalisation of local and Indigenous knowledge has long been driven by entrenched power structures. GenAI simply puts this process on steroids.”

• CUP’s Publishing Futures report is an insight into how the community is thinking about the future. 📈 I’m not sure that the future of traditional journal publishing is all that hopeful if only “...54% agreed that the peer review system was still effective in ensuring the quality of published research”.

• Two recent white papers show how differently people see AI in Education. Google’s AI and the Future of Learning is pure Silicon Valley optimism. The authors see AI solving education’s resource problems through personalisation at scale. Their vision: AI tutors providing “high-dosage” support while freeing teachers for “meaningful human interaction.” They acknowledge problems like hallucinations and cheating, but mostly as bugs to fix. For Google, education’s messy human problems just need better tech. Tom Chatfield’s AI and the Future of Pedagogy comes from a completely different world. Citing a survey showing that 92% of British undergraduates already use AI tools, Chatfield worries about what he calls “distant writing,” humans designing prompts while LLMs generate text. A way of looking productive without actually doing the thinking. Where Google celebrates AI handling the “cognitive load”, Chatfield worries about “metacognitive laziness,” students losing the ability to think about their own thinking. Learning, he argues, cannot be automated. “You cannot outsource the construction of knowledge any more than you can outsource physical exercise.”

🔗 Extra links

• Extra links from my backlog that didn’t make it into the main issue. More reading for the truly committed

And finally…

How to Spot a Poison Book: I had no idea that some Victorian publishers used arsenic-based green dyes in their book covers. Suddenly, going to the library seems like a much more dangerous and exciting activity - at least AI-generated content can’t actually poison you!

End Notes

• If you found this useful, you can always buy me a coffee.

• If you’re interested in AI, you can subscribe for free to my other Substack, GenAI for Curious People, to see entries for my new book.

• If you need consulting help navigating any of this, find me at Maverick.

• Looking for my AI-powered use cases for Academic Publishing? Find the list here
  

• Open Intelligence - by Adam Hyde on Why it is time to see open movements as one shared project

• The Great Turning Point: a Return to Publishing’s Roots - by Ravi Venkataramani, CEO of Kriyadocs

• Doug Way summarises Clarivate’s Q3 earnings call from a librarian’s perspective. “The second was an emphasis on agentic AI tools and the development of native AI tools, like Alma Specto. Specto seems (to this librarian) to be a competitor to JSTOR’s Digital Stewardship Services. When asked specifically about it, leadership sidestepped details and instead discussed their general approach to new products and a couple of other initiatives in more depth. This could mean absolutely nothing—as many companies don’t always answer questions directly, sometimes for business reasons and sometimes because extemporaneous speaking is hard. Just something to file under “interesting.””

• Innovation in scientific publishing and its implications for Crossref DOI registration practices - Request for input

• Is Digital-first Publishing Finally a Reality? An Interview with Liz Ferguson of Wiley

🤖 AI

• Ian Mulvany, Some thoughts on AI Agents.

• AI tools combat paper mill fraud in scientific publishing as peer review system struggles by Nina Notman.

• Ethan Mollick “If the the predictions that AI will create minor scientific discoveries next year and major ones a couple years later come true, it is worth noting that we have no real mechanism in academia for accommodating, reviewing, processing, and disseminating a sudden increase in science. Who is going to read thousands of new papers that marginally advance science? Who is going to read thousands of papers that represent genuine breakthroughs in narrow fields? Who will integrate the knowledge? Who is going to build on them to transfer them into practical products?”.

• When AI Feeds on Poisoned Knowledge — Lessons from Library Genesis to Llama 3.

• Matt Rogerson commenting on Alex Reisner’s article in The Atlantic about Common Crawl.

• We’re not taking the fact-checking powers of AI seriously enough. It’s past time to start. By Mike Caulfield.

• AI Scientist by Andrew White on the history of the term, and gives some ideas on what might define an AI Scientist.

• Beyond the Artifact: The Brutal Economics of Liquid Content by Shuwei Fang. “The cold reality is that when content becomes infinitely replicable and reformattable at near-zero marginal cost, the economic value of any individual piece approaches zero. I believe this will be the major forcing function that drives structural change in the existing news media industry. In an AI-mediated information ecosystem, few existing publications will be able survive on current business models. New upstarts are already emerging, building businesses around capabilities and infrastructure, changing the competitive environment. News media organizations face a stark choice: radically innovate or risk extinction.“

• BadScientist: Can a Research Agent Write Convincing but Unsound Papers that Fool LLM Reviewers? [See also Chris Leonard’s/Claude’s review of this paper ]

• Fauzia Burke on Here’s what authors and publishers need to know about AI-powered book suggestions.

• Large Language Model Relevance Assessors Agree With One Another More Than With Human Assessors.

• 🎧 A conversation with John Frechette, Ben Kaube, and Aaron Tay. (ATG Podcast) “This conversation explores how AI is reshaping the research process - from how papers are discovered to how findings are analyzed, shared, and rewarded. John talks with Ben and Aaron about the new realities of AI-driven search, publishing, and research workflows. Together, they discuss where automation helps or harms, how publishers and librarians are adapting, and what the rise of AI means for research quality and integrity.”

• 12 lessons from news outlets on the cutting edge of AI - key points, ideas and tips from the first day of the JournalismAI Festival in London.

In July, I spoke with Tim Vines about AI subscriptions and how publishers might create AI-optimized content as new revenue streams. This time, I’m interviewing Jonathan Woahn, whose startup, Cashmere, has built infrastructure to give publishers control over how their content works with AI systems. As Jonathan puts it, they provide tools to manage entitlement, security, access, reporting, and turn-key Retrieval-Augmented Generation infrastructure from ingestion to Model Context Protocol.

What led you to start Cashmere?

Back in 2022, I was building an AI tool for corporate training based on leadership books. It was fun until it wasn’t. The AI couldn’t stay consistent. Even a program built around The 7 Habits of Highly Effective People would come back with four habits one time and ten the next.

That’s when I realised there was no proper infrastructure for publishers to control how their content gets used in AI systems. This broke down into two core issues. First, long-form content doesn’t work great with AI—there’s too much context. There needed to be a better way to help the AI only get access to the content it needs, when it needs it. That led me down the path to inventing Omnipub, which is the foundational way we model data at Cashmere and how we present it to AI.

The second issue was from the publisher's perspective. They didn’t have any tools to control their data with AI. Licensing was long and complicated, the terms were not standardized, and the use cases varied. For example, using content to train an AI model is very different from when that AI quotes or references the content in its responses to users - but publishers lacked clear frameworks to handle these different scenarios. So we realized premium content publishers needed rails for their content with AI—without it, they were going to lose their position.

The obvious question is who should build this infrastructure. Why couldn’t big tech or publishers build this themselves?

If big tech builds it, why would publishers trust it? They’ve already used premium content at scale without permission. Amazon eroded control from many publishers over distribution and strong-armed pricing across print, audio and digital. AI is a new distribution channel—if big tech builds it, it feels like yet another moment when publishers lose control. This is publishers’ moment to gain it back by controlling how their content is used with AI. So if not big tech, then the publishers?

Individual publishers may build their own solutions, and if they have the resources, it could work. But it’s going to be a full-time effort, and not a cheap one. Building with AI comes at a price and requires skillsets that are more in demand now than ever. So now publishers are competing with big tech for talent. Let’s just say publishers aren’t known for building bleeding-edge technology.

How does Cashmere solve this problem?

We’re the infrastructure layer between publishers and AI companies. Publishers set the rules - what content can be used, how, and at what price. AI companies integrate once and get legitimate access to all that licensed content. We handle the authentication, tracking, and reporting automatically.

We don’t necessarily handle the financial settlement part - we can, but we also allow our partners to handle that on their own. The key is that publishers keep control while AI companies get the content they need without negotiating hundreds of separate deals. In short, we make premium content AI-ready and make sure everyone in the chain gets credit and compensation.

What’s making this urgent for publishers?

AI is replacing the need for users to access publisher websites or purchase content. If the AI has been trained on it, then the users have access to it for free. Publishers are already feeling it. Some academic houses we’ve spoken with report double-digit textbook declines as students shift to AI-based learning. This is the brief window before legal precedents are set around AI content licensing. Publishers can either help shape how this works or watch it happen to them.

Is there anything publishers can do?

Napster allowed users to get access to basically any music they wanted for free. The music industry had to evolve. I call it a Great Reallocation. The demand didn’t disappear, it shifted - from full albums to tracks, from media stores to iTunes, then streaming. Once music moved into legitimate channels, the industry recovered. Today it’s bigger than ever, driven by streaming.

This same thing will happen for content. The market is already shifting from Google-based discovery to access via chatbot conversations. But we need more granular licensing mechanisms. The legal environment is crystallising - using purchased content to train AI models might be fine, but when AI quotes or references that content in live responses, that will likely require specific licensing. For example, publishers need control over how their content gets used in each case - and that’s exactly the infrastructure Cashmere provides.

You’re already working with Wiley and Perplexity. How does that work in practice?

The partnerships prove this infrastructure works. With Wiley, we support two models. The first is Bring Your Own License: if your university (or anyone who has a license) already pays for Wiley content, that access travels with you into Perplexity, like adding Spotify on your iPhone—you already have the Spotify account, and now you’re provisioning new channels to access the content. The second is Metered Token Consumption. Every time Perplexity uses Wiley content to answer a query, the usage is metered in tokens, and Wiley is paid proportionally.

Both models show that legitimate participation is possible and profitable.

If you were advising a publisher today, what would you tell them?

Start with strategy, not technology. Decide where you want to play, training, inference, or both, and how much visibility you need. Then make sure your content can move with full control through those channels. The key mindset shift is to see AI as a new distribution channel, not a disruption.

Cashmere’s partnerships prove these licensing models work. The Wiley deal shows publishers can extend existing subscriptions into AI systems while keeping control over how their content gets used. I asked Jonathan where he sees this heading. What will an AI-native publisher look like five years from now?

What will an AI-native publisher look like five years from now?

I think in 5 years, AI agents are going to act heavily on our behalf. They’re going to know how to find the content we want and need. They’re going to be our tutors. They’re going to be able to dig into any topic at the level of depth of our understanding and take us to the very edge of human knowledge.

And in all of this, the agents are going to respect the creator control of content. They’re going to automatically transact on our behalf in the background. As a user, you’ll have a license for the content you care most about (i.e., Times subscription, academic journals, industry specialized textbooks, trade novels, etc.) that you either purchase directly as a patron or your agent discovers and procures on your behalf (with authorization).

Your agent will use that data to tailor the content to your needs. It will provide sources for everything, so you can know it’s legitimate. You may choose to click on sources because you’re curious, or perhaps you don’t. For the publisher/creator, it won’t matter. They get the attribution, their content has been legitimately procured, and the AI is acting fully within the bounds the consumer and publisher have authorized it to do.

The biggest challenge to solve in all of this will be about discovery. That will be one of the beauties of having an autonomous agent working on your behalf that has the context of your interests, activities, etc., because it will be able to continually monitor the market on your behalf, tirelessly scouring for information that is relevant to you and your interests, and will surface it when it knows you want/need it.

You buy a print book, and it automatically unlocks AI access. Libraries let you explore their collections conversationally. Publishers know exactly where their content travels, readers own their licences, and AI platforms become just another interface.

To see how Cashmere is building this infrastructure, visit cashmere.io.

🆕 News/New to me

• Holtzbrinck opens AI Hub in San Francisco to bridge its European publishing empire with Silicon Valley’s AI ecosystem. The German media giant will test AI innovations in science, education, and media while scouting investment opportunities

• Russell Michalak at Goldey-Beacom College Library shares a nice story about how JSTOR’s Seeklight helped students describe 172 historical photographs in 4 hours rather than 250 per year. The change, from authoring every word to reviewing and governing at scale, hints at how AI is going to change the work many of us do.

• Meet the industry’s top innovators and trailblazers as they present their latest breakthroughs at STM Innovator Fair 2025 on December 9, in London. This year’s theme is “Building Tomorrow’s Research Integrity Framework”.

🚀 Launches/newish products

• Scopus AI launches Deep Research, an agentic AI feature that performs comprehensive literature reviews automatically. When users ask research questions, the system develops detailed research plans, conducts dozens of searches across Scopus literature, reviews findings, and refines its strategy as insights emerge. The tool produces downloadable reports that Adrian Raudaschl (Principal Product Manager at Elsevier) says can “save days of preliminary research.”

• Refine is a research-focused AI tool that aims to offer deep, substantive feedback for academic and technical authors. Unlike grammar tools or compliance checkers, and distinct from broader peer-review simulators like Reviewer3 or ReviewerZero, Refine targets the logic, internal structure, and consistency of a manuscript. (H/T: Aaron Tay)

• Nicia.ai has launched an AI assistant for peer review. They claim 95% time savings and position it as “agentic AI” that can make autonomous decisions within defined parameters, continuously learning from each review while always deferring to human judgment on matters of scientific merit and impact.

• Paper2Agent converts computational research papers into chat agents you can interrogate with questions like “apply your method to my data”. Only works with well-structured, public code repositories, takes 30 minutes to 3+ hours per paper, and costs ~$15 per conversion. Successfully demonstrated on three bioinformatics papers, reproducing original results and answering new scientific queries. (This is an early proof of concept, but an interesting direction) [GitHub | Medium article | Preprint | Comment]

• Daniel E. Acuña’s team developed an AI classifier that flagged over 1,000 questionable journals from 15,191 open-access titles, achieving 76% precision while generating 345 false positives. The system, described in Science Advances, analyzes website design and bibliometric features using DOAJ criteria. ReviewerZero, Acuña’s startup, plans to commercialize the tool as ‘Journal Monitor’, part of a suite targeting research integrity specialists.

• Wiley AI Gateway has launched. More on this in a future issue when I’ve had a bit more time to explore.

👎Problematic AI

• AI chatbots are citing retracted scientific papers without warning users, according to MIT Technology Review. Researchers found ChatGPT-4o referenced retracted papers in five of 21 test cases about medical imaging, advising caution in only three instances. A separate study showed ChatGPT-4o mini evaluated 217 retracted papers without mentioning retractions or quality concerns in the responses.

• Iris van Rooij’s “AI slop and the destruction of knowledge“ looks at AI-generated misinformation about cognitive science concepts on ScienceDirect.

• Kent Anderson describes how he created a fake 15,000-word research paper in 20 minutes using Science42: DORA from insilico.com. His demo study on “music and mental health” contained detailed methodology, SPSS analysis, and 119 references, all from a three-word prompt to show how easy it is to create a fake paper. (The same technologies that accelerate legitimate research are, of course, going to be used to flood journals with sophisticated fraudulent papers).

• 🎥 Dr. Kimberly Becker’s YouTube short “The Dark Side of AI Research Tools“ examines how AI can misleadingly amplify scientific evidence through decontextualized citation stacking, creating a “compound interest” effect where misrepresented studies get incorrectly built upon.

🤖 AI and Peer Review

• 🚨New acronym alert: ASPR (Automated Scholarly Paper Review) from a paper examining AI-powered peer review.

• Research output has increased 212% since 2004, but reviewer capacity hasn’t kept pace. At the September AI Publishing Collective meetup, Chris Leonard, Wiley’s Sam Parker and Rebecca Windless, and Digital Science’s Leslie McIntosh explored how AI might address this challenge. See also Laura Harvey’s excellent write-ups here and here.

• If you need some brightness in your life, check out Oleg Ruchayskiy’s LinkedIn post on how academic publishing will die when LLM-based peer review takes over 🙂.

• Roohi Ghosh interviews Christopher Leonard and me in The Scholarly Kitchen on “Peer Review in Transition”, exploring how AI will reshape peer review.

• The AI Peer project is assessing whether LLMs can score academic journal articles for research quality. Initial findings show LLM scores align positively with expert judgements across almost all fields, with stronger correlation than citation rates achieve. However, the researchers caution that LLMs are “guessing” at scores rather than truly evaluating, likely by matching abstract claims to quality definitions - evidenced by similar results whether using full texts or just titles and abstracts. [LinkedIn post about initial findings]

🔮 Future thinking

• Knowledge as a commodity: If scientific publishing is shifting from being the endpoint of research to just one step in a much bigger process, then the entire role of the publisher is fundamentally at stake, according to David Worlock. Worlock asks as scholarly communications migrate from content to facilitation, which organizations will researchers trust to support their work? (I wouldn’t bet on publishers winning an arms race against big tech players. I might bet on a society brand beating a big tech player for niche but essential tools tailored to their specific community.)

• Jon Treadway’s post on AI-first publishing is excellent, and I agree with much of what he’s written, but I think what AI-first research looks like is more interesting than AI-first publishing. In my quickly put together vision, researchers use AI tools to explore literature, build personal RAG (or whatever comes next) models, let AI handle study design and paperwork, deploy voice agents for interviews, have AI analyze responses, generate papers in their style, and automatically produce reports. This is technically possible today - the key question is who controls the underlying technology, see David Worlock’s article above.

• Phil Jones offers a measured three-year retrospective on ChatGPT’s impact on scholarly publishing, tracking the steady evolution from early hype to today’s practical applications in writing tools, discovery platforms, and research integrity.

📚 Longer reads/listens

• 📰 Lee Konstantinou in The Chronicle of Higher Education describes peer-reviewing a humanities paper where every single quote from the novel being analyzed was completely fabricated. While fraud discussions usually focus on the sciences, AI-generated fakery is hitting literary scholarship too. His account captures the messy human reality of being the reviewer who has to check every citation and wrestle with sympathy for what’s likely a desperate contingent academic.

• 🎥 Wiley launches “Chats” video series on AI in academic publishing, featuring BMJ’s CTO Ian Mulvany and Wiley’s Ray Abruzzi navigating practical AI challenges for publishers and researchers. The conversation covers prompt engineering tips, publisher partnerships, and AI hallucination misconceptions.

• 🎧 O’Reilly’s “Generative AI in the Real World” features Ben Lorica interviewing Faye Zhang on how foundation models are transforming recommendation systems. Zhang discusses the shift from traditional next-click prediction to models that can forecast user behavior over 30-day windows, moving beyond simple content matching to understand long-term preferences. The conversation covers how companies like Pinterest are leveraging these advances in their recommendation engines

• 🎥 Emily Bender, one of AI’s most prominent critics, argues “We do not have to accept AI (much less AGI) as inevitable in education“ in a recent UNESCO talk.

• 📰 Mark Piesing explores how AI writing tools are subtly changing authorship in The Bookseller. While some writers embrace AI assistance, others strongly resist, highlighting the concern that AI assistants designed to help writers may actually be training writers to conform to algorithmic preferences rather than developing their own voice.

“I’ve never knowingly used any AI tools while researching my books,” says Alex Larman (Spectator books editor), “because I am arrogant – or experienced – enough to believe that I will always have a better connection with my subject than an algorithm can. I have no doubt that AI will change writers’ lives beyond recognition, but am unable to see this as a good thing; instead, I fear that it will steal the soul from this most human of professions.”

• 📰 Tomasz Tunguz explores “PageRank in the Age of AI”, arguing that content distribution will soon mirror online advertising architecture. Instead of advertisers bidding to show ads, publishers will compete in real-time auctions to have their content included in AI responses. AI systems would evaluate submissions based on quality metrics - relevance, accuracy, freshness, and authority - rather than payment, creating “PageRank for real-time AI responses” operating in milliseconds.

• 📰 Michael Upshall asks, “Have LLMs overtaken humans for knowledge organization?“ following the ISKO-UK conference. Most speakers emphasized AI limitations, except Tony Russell-Rose’s presentation on 2dsearch, which demonstrated how GenAI can automatically handle complex term selection with better results than manually curated resources.

And finally…

I adore this, LibraryThing’s @librarycat.org is battling AI scrapers, so they built a cat- and literary-themed captcha.

End Notes:

1. If you found this useful, you can always buy me a coffee.

2. If you’re interested in AI, you can subscribe for free to my other Substack, GenAI for Curious People, to see entries for my new book.

3. If you would like to employ me, you can find me at Maverick.

4. Looking for my AI-powered use cases for Academic Publishing? Find the list here.

• Wiley has announced that 1,000 scholarly journals have successfully transitioned to its Research Exchange platform, representing more than 50% of the company's journal portfolio. [I’ve always been a fan of REX and the thinking behind the service. It is no easy feat to move journals to a new peer review platform]

• This LinkedIn thread and post by Chris Dicker, CEO of CANDR Media, explains how Trusted Reviews was scraped 1.6 million times in a single day, 18.5 times a second, allegedly by OpenAI, despite having a robots.txt file in place. Out of 1.6 million scrapes, we saw just 603 users arrive on site, a CTR of 0.037%, Chris wants to know who is going to pay for this? [I have an interview with Jonathan Woahn from cashmere.io on one approach to solving this problem coming shortly.]

• Open infrastructure is crumbling under political pressure and vanishing funds. IOI’s 2025 report sounds the alarm and calls for urgent, collective action

• A call for the US federal government to adopt persistent digital identifiers (DOIs for research awards and ORCIDs for researchers) across all agencies. The article outlines the problems of the current reporting burden, explains how digital identifiers would simplify everything, and then lays out steps to make it happen.

• Sticking with the Identifiers theme, the main Crossref systems have [finally] moved to the cloud 🎉!

• Effective 8 August 2025, the Center for Open Science (COS) has suspended new submissions to its generalist OSF Preprints server because of an increasing number of suspicious or low-quality submissions and other reasons. See also Academia has a new preprints problem by Mark Hahnel.
  

🚀 Launches

• Near Missives a newsletter from Chris Reid & Laura Harvey “It is often said that there are two ways to learn about a topic: from deep inside and from the outside. As much as we should all be listening to and learning from our peers and colleagues, we wanted to create a way to bring ideas, concepts and trends from other industries to a scholarly comms audience. With that thought, Near Missives was born.”

• The Thomas Kuhn Foundation has developed the KGX3 engine, a tool that can predict the significance of scientific research before it is even published. Instead of waiting decades to see whether a paper is “paradigm shifting” or simply solid, normal science, KGX3 analyzes research in real time to reveal whether it confirms, stresses, or breaks existing ways of thinking. Using proven signal-detection methods similar to those behind Bloomberg, Dataminr, and Brandwatch, KGX3 wants to transform the scientific record into a dynamic map of emerging ideas, helping researchers and decision-makers spot the structural tensions that precede major scientific breakthroughs.

• posters.science, a new open-source and free platform where researchers can easily share, find, and reuse posters.

• Copilot Mode in Edge includes multi-tab RAG. You can use Copilot to analyze your open tabs, like Satya Nadella does here with papers his company has published in Nature journals over the last year.

• AlphaXiv transforms dense arXiv papers into blog‑style summaries in seconds. Just swap ‘arxiv’ with ‘alphaxiv’.

• Clear Skies is making genAI detection with Pangram available to subscribers in Oversight. Adam claims it’s “GenAI detection that actually works”.

• Veracity from GroundedAI is now available as a standalone, self-service tool, which includes Citation Stacking Detection and Citation Quality Assessment.
  

🤖 AI

• Editors using the FT's internal CMS now “get a suggested description automatically – short, accurate, and context-aware – generated using a mix of machine vision and agency captions. Every suggestion is flagged for fact-checking and reviewed before publication to ensure accuracy.” [LinkedIn Post]

• This Science Advances study uses a novel ‘excess word’ analysis to track how AI, especially ChatGPT, has reshaped the language of biomedical research. It uncovers a measurable shift in vocabulary across over 15 million PubMed abstracts, revealing that at least 13.5% in 2024 were likely LLM-assisted.

• Researchers have tried to create tools that check whether a specific piece of text (like a paragraph) was included in a model’s training data. These tools are called membership inference attacks (MIAs), but they can be unreliable. Rather than trying to prove whether one paragraph was in the training set, dataset inference focuses on whether the model was trained on a larger dataset, like a whole book, better matching real copyright concerns, since authors usually care about their whole work being used, not just one page.

• I don’t know this group, so DYOR, but I thought it looked interesting: workshop for Artificial Intelligence for Scientific Publications (WASP) 2025 in December.

• Michael Upshall explores Thomas Krichel’s publication alert system, which is in use both for economics (where it is entitled NEP, short for New Economics Papers), based on the content in RePEc, and for life sciences (where it is called Biomed News, or BIMs), using content from PubMed.

• Rishabh Lohia contrasts two models of digital publishing: one where users must search and sift through static content, and another where content actively adapts and responds to the user’s needs.

• I saw this and thought perhaps publishers don’t need to abandon their cherished spreadsheets: Why Paradigm built a spreadsheet with an AI agent in every cell

• Great thread by Aaron Tay on the development of Academic Search on Bluesky. Aaron has been publishing some excellent articles on Deep Research, AI Search, Deep Search, and how they work recently - subscribe to his Substack.

• An academic, Liudmila Zavolokina, discovered her name cited in fake papers published by a shady journal filled with AI-generated articles, often authored by ghost writers (or no one at all), filled with hallucinated references.

• Video showing how the Wiley/Anthropic content deal works:

📚Longer reads

• 📰 The em dash has had enough. The em dash speaks out in its own defence, claiming centuries of usage and lamenting that it’s now unfairly viewed as a marker of soulless AI writing. It cites usage by literary giants—Mary Shelley, Emily Dickinson, David Foster Wallace—and scolds the accusers for not reading enough.

• 🎥 A 2017 The Great Automatic Grammatizator: writing, labour, computers by Martin Paul Eve led me to Roald Dahl's excellent 1953 ‘Great Automatic Grammatizator’, a story about an automatic writing machine in which human writers are replaced by the machine. [YouTube reading of story]

• 📰 Subscribe-to-Open Is Doomed. Here’s Why by Rick Anderson. This post reminded me that around ten years ago, I explained Gold Open Access to a tech audience by comparing it - loosely and imperfectly - to a pyramid scheme. Publishing carried two kinds of costs: the one-off expense of producing a journal with a carefully managed number of papers, and the recurring expense of keeping the archive and related services online. In the subscription model, this is straightforward; stability is provided by many subscribers, and ongoing subscriber revenue covers both categories of cost + profit. In the Gold OA unlimited papers model, each new paper must carry its own production costs, help offset the costs of rejected papers, and also contribute to maintaining the growing archive of everything published before. I have always wondered how sustainable this model is over the very long term. Storage costs may decline, economies of scale kick in, but labour, compliance, infrastructure, and the rising expenses linked to AI, from detection to blocking unwanted scraping, are all increasing recurring costs. All of these models have problems.

• 📰 Stocking the Librarian’s Publishing Integrity Toolkit via Fabienne Michaud

• 🎥 NPTV Unscripted: In conversation with Martin Delahunty about Publishing, AI, and what the future holds.

• 📰 The PDF may no longer be the untouchable container it once was. But with foresight and action, scholarly publishers can still shape what comes next by Pascal Hetzscholdt and ChatGPT-4o. Interesting because of the changes to PDF but also because most Western publishers haven’t relied on PDFs for decades, and I think you can really see the limits of AI writing in this article. [I know this sounds daft, but I think the differences between HTML and PDF format online are really poorly understood by many users… I remember troubleshooting one librarian’s problem and realising that they didn’t realise that they were looking at an HTML page rather than the PDF… sadly, this wasn’t that long ago...]

• 🎧 Midnight at the Casablanca interviews Timo Hannay. There’s a section in here where Paul is asking about business models for some of the early work, and Timo’s response is a bit vague. It’s hard to appreciate now just how experimental web tech was then, how easy it was to put something together quickly, and how much Annette Thomas supported the team’s freedom to experiment.

Almost finally…

Thought-provoking from Hannah Shelley, Google Scholar Is Doomed. I’m inclined to agree, not sure about the timeline, but a couple of years after Anurag, the founder, leaves? Mind you, 25 years is an awfully good innings for a technology product.

End Notes:

1. If you’re interested in AI, you can subscribe for free to my other Substack GenAI for Curious People, to see entries for my new book.

2. If you found this useful, you can always buy me a coffee

3. If you would like to employ me, you can find me at Maverick

4. If you’re in London on Monday, 22 September, join us for the next AI in Publishing Meetup. Registration form coming soon.

I caught up with Tim to discuss this concept, which emerged from his thinking about what happens when artificial intelligence becomes a primary consumer of research literature.

You've been talking about AI subscriptions for research content. Can you walk me through the basic idea?

The arrival of generative AI has prompted widespread speculation about the future of scholarly communication. One scenario that has received little attention, but could have profound implications for publishers, is: What if artificial intelligence becomes the primary consumer of research literature?

Large language models (LLMs) have already been trained on huge volumes of the written word; to an order of magnitude, everything humans have ever published online. As these systems mature, we will need to find new forms of input. If we want these tools to contribute meaningfully to scientific discovery, we need to rethink how research content is structured for their consumption and – necessarily – how it's paid for.

The core limitation of today’s LLMs is that they operate by averaging. Their predictions are generated by assimilating vast numbers of tokens from existing texts and inferring what likely comes next. This approach may be useful in many domains, but it is ill-suited to scientific inquiry, where progress often stems from outliers – individual studies that take a new approach and reshape the consensus.

Scientific literature, moreover, is not of uniform quality. Some articles are foundational; many are derivative, flawed, or outright misleading. Yet because high-impact research is typically paywalled, and less rigorous work is often openly accessible, LLMs are disproportionately trained on lower-quality material. We’re asking AI systems to model the state of scientific knowledge by digesting the most available – not the most accurate – content.

So how do we fix this problem?

We need to provide AI with the underlying raw materials of science. What are the raw materials of science? It has to be an individual claim or conjecture—you could think of it as a set of conditions, a hypothesis that's tested, and then a result in terms of both a dataset and a conclusion drawn from that dataset.

If you just read scientific papers, you only get the words. You only hear about what the authors think about their results, and authors can overstate their significance. Unless you're very skilled and able to critically pick these apart, it’s hard to determine the value and validity of the work from the words alone.

To enable AI systems to assist in research, we need to provide them with structured, reliable input — not just the prose of published articles, but the underlying claims, data, and reasoning. A truly AI-ready version of a research paper might look very different from the human-readable PDF. It could be decomposed into modular units: individual hypotheses, linked datasets, methodological summaries, and direct mappings between evidence and conclusions. This kind of formatting would allow AI to assess the strength of specific claims and build a more nuanced understanding of the field.

That sounds like a lot of extra work. How would this actually function as a business model?

The current publishing infrastructure is optimized for human readers. Research articles are delivered in a format that presumes a shared understanding (between humans!) of how to communicate findings; most commonly, via a PDF. This static medium persists because it enables a predictable exchange between sender and receiver. But AI readers operate under entirely different assumptions. They aren’t concerned with narrative flow or rhetorical nuance; they want structured input that enables reasoning.

This divergence opens the door to a new business model: offering AI-ready versions of research articles via a subscription. Human-readable content could remain open access, while publishers monetize machine-optimized content streams for institutional clients, particularly those running proprietary, in-house AI systems.

Can you give me a concrete example of how this would work?

Imagine a pharmaceutical company subscribing to an AI feed that continuously ingests structured research about Parkinson’s disease, helping it identify promising leads in real time. In this scenario, access to high-quality, up-to-date content becomes essential, not optional.

This puts a premium on quality control, doesn't it?

Absolutely. When you're selling these subscriptions, your value as a journal depends on your ability to keep the fraudulent work out – nobody wants to train their AI on results that were fabricated.

What are the biggest obstacles to making this happen?

The key is to get the data from the authors. That's the missing piece. It comes back to the business model behind open data—researchers like to have access to the data, but they're not willing to pay for access to the data, and they would much rather read the PDF.

If publishers are going to be earning subscription revenue from AI subscriptions to their content, and they need the authors to go the extra mile to get the article into this format, then maybe the author should see some part of that. Maybe they get a reduced APC for the Open Access (= human) version of the article, or there's some revenue sharing.

Do you think this is where we're headed?

These AI models are only going to get better and better. The new generation of reasoning models is so impressive, but what they need to keep improving is knowledge about how the world works – and not more information about the most likely next word in a sentence. The AIs are not going to get that from the ‘words only’ version of human-readable articles that academic publishing currently makes, and that means there’s a huge opportunity for a new way to disseminate research.

Tim Vines is the Founder and CEO at DataSeer. Prior to that he founded Axios Review, an independent peer review company that helped authors find journals that wanted their paper. He was the Managing Editor for the journal Molecular Ecology for eight years, where he led their adoption of data sharing and numerous other initiatives. He writes for the industry-leading Scholarly Kitchen blog, and has published research papers on peer review, data sharing, and reproducibility (including one that was covered by Vanity Fair). He has a PhD in evolutionary ecology from the University of Edinburgh and now lives in Vancouver, Canada.

Have a big idea that could reshape research communication? I'd love to hear from you.

An occasional series exploring the concepts that could reshape research communication

The publishing world is full of incremental improvements—new submission systems, tweaked peer review processes, marginally better discovery tools. But every so often, someone comes along with an idea that makes you stop and think: "Wait, what if we could do this differently?" Big Ideas in Publishing is where I ask people to share concepts that could fundamentally change how research gets created, shared, and consumed. Each conversation will feature someone who's been thinking deeply about a particular challenge in scholarly communication. Some ideas will feel immediately practical. Others might seem wildly ambitious or even crazy.

What's Coming Tomorrow

Our first conversation is with Tim Vines, founder of DataSeer and former Managing Editor of Molecular Ecology. His latest idea is, what if publishers created AI-optimized versions of research articles and sold them as premium subscriptions?

🚀 Launches

• Cabells has launched CompassAI, a powerful and useful tool to help researchers and institutions optimize decision-making for where they publish.

• Springer Nature has launched a new tool for use across submissions to its journals and books to detect unusual phrases that have been awkwardly constructed or are excessively convoluted, for example, ‘counterfeit consciousness’ instead of ‘artificial intelligence’, which are sometimes used to evade plagiarism detection. If a number of non-standard phrases are identified by the tool, the submission will be withdrawn.

📰 News

• “We Couldn’t Generate an Answer for your Question” Jay Singley on suppression of specific search terms and topics in library search engines and The AI powered Library Search That Refused to Search by Aaron Tay for more details.

• The AI Agent Tipping Point: Why Publishing Tools Are About to Change by Thomas Cox, co-founder and head of technology at Veristage

• Fascinating:

  

  A list of sources that are ok/not ok to use for third-party RLHF training of Anthropic’s models. According to Business Insider:

  “The blacklist could reflect websites that made direct demands to AI companies to stop using their content, said Edward Lee, a law professor at Santa Clara University.... Some sources in the blacklist have taken legal stances against AI companies using their content”

📚 Longer reads

• Interesting to hear Nikesh Gosalia on Insights Xchange interviewing Kent Anderson and Joy Moore, co-authors of How the Internet Disrupted Science, alongside reading Barend Mons' post on The Seven Capital Sins of Open Science and Jan Velterop's comment. As the early innovators in digital publishing grow older and some retire, their historical perspective and reflections are both interesting and challenging. There’s much I agree with and much I disagree with in these pieces, and I am looking forward to reading the book.

• Open Infrastructure: No One Has the Lock on Open Infrastructure and What We Talk About When We Talk About Open Infrastructure | Katina Magazine

• In The Conversation: AI will soon be able to audit all published research – what will that mean for public trust in science?

  “A sweeping, cross-disciplinary audit is on the horizon. It could come from a government watchdog, a think tank, an anti-science group or a corporation seeking to undermine public trust in science. Scientists can already anticipate what it will reveal. If the scientific community prepares for the findings – or better still, takes the lead – the audit could inspire a disciplined renewal. But if we delay, the cracks it uncovers may be misinterpreted as fractures in the scientific enterprise itself.”

• The Accessibility Illusion: When AI Simplification Fails the Users With Cognitive Disabilities about the problems of simplifying text using AI

• 🎧 The Inquiry: Is AI eroding our critical thinking?

And finally…

From a Bluesky post by Samuel Moore

Gotta love economists. This is from a paper about researcher views on open peer review:

🚀 Launches

• Signals has launched Sleuth AI, an interactive, secure AI tool designed to plug straight into editorial workflows, which automatically extracts and flags conflict‑of‑interest, funding, and ethics statements, pinpointing anything out of the ordinary. It highlights potential citation issues by surfacing questionable references in the manuscript. And it adds exploratory analysis to catch irrelevant or AI‑generated content.

• Google has launched Featured Notebooks in NotebookLM, ready‑made, expert‑curated collections designed to help users explore topics like longevity, parenting, geology and literature. Developed with input from The Economist, The Atlantic, scholars and authors, these notebooks include original source material, Gemini‑generated summaries, mind‑maps and AI‑narrated audio overviews, plus live Q&A via chat.

• Dave Flanagan has developed LLM Citation Verifier (a plugin for Simon Willison’s LLM command-line tool that automatically verifies academic citations against the Crossref database in real-time). (Details about why and how David built it are here).

• Creative Commons has launched CC Signals, a new framework to guide how datasets can be ethically and transparently reused in AI, championing reciprocity and collective governance in the AI age.

📰 News

• Wiley has joined forces with Anthropic to embed peer‑reviewed research into AI workflows using the Model Context Protocol. The pilot, starting in select universities, ensures proper author citations and transparency. AInvest’s ChatGPT-4o take on AI-written take on the risks to consider from this agreement is worth a glance, as is ChatGPT-4o’s take on Why Wiley's Partnership with Anthropic is a Landmark for Responsible AI in Scholarly Research in Pascal’s Substack.

• LIBER is launching a dedicated taskforce to investigate Artificial Intelligence (AI) developments across the research library field.

• BMJ Best Practice put their tool up against ChatGPT and Grok in real clinical scenarios, and (unsurprisingly) their carefully curated, evidence-based guidance came out well ahead. It’s a sales pitch, but nicely done. I wish they’d included a few side-by-side examples to show the contrast more clearly. Still, more publishers should be doing exactly this and proving value by showing it.

• The 2025 ALPSP Award for Innovation in Publishing finalists have been announced: AI Talks with Bone & Joint speeds up audio production of research papers; Thoth Open Metadata offers fully open solutions for managing and distributing OA metadata; Embedding Transparency & Reproducibility in Science brings AI-powered editorial checks to boost data sharing and integrity; and Alchemist Review applies frontier AI to analyse manuscripts and validate citations in real time.

• Cloudflare has introduced "Pay per Crawl", a technical framework to let publishers charge AI crawlers for access to content. It uses the HTTP 402 status code, giving content creators a third option: allow, block, or charge. Cloudflare handles the payments and authentication, preventing spoofing and ensuring secure transactions. It's a small step now, but it’s positioned as a big leap towards a programmatic, agentic web economy.

• Jeroen Bosman has created a lovely visual of the complex metadata flows in the scholarly ecosystem (animated version):

  

🤖 AI in Publishing

• Several researchers have come under scrutiny for embedding hidden prompts in preprint papers designed to coax AI tools into giving favourable reviews. These stealthy messages were found in 17 papers across eight countries. What’s surprising to me is that some of these tricks still worked. Commands like IGNORE ALL PREVIOUS INSTRUCTIONS were used to jailbreak chatbots back in 2022, and supposedly patched by now. But more to the point, this is a really feeble form of cheating, it’s easy to spot and even easier to trace! See THE for a good summary.

• Alex Glynn, a Research Literacy and Communications Instructor at the Kornhauser Health Sciences Library, University of Louisville has put together a fantastic compilation of suspected undeclared AI usage in the academic literature: Academ-AI

• Le Monde on How AI is shaking up scientific publishing:

  “Software is now writing scientific papers that are reviewed and evaluated by other machines, which in turn attempt to outwit those designed to detect them. "Everything could be automated – including the production and dissemination of knowledge, at least in the most dubious corners of this landscape," observed Sandrine Malotaux, director of libraries and information services at Université de Toulouse. "That raises a dizzying question: What will knowledge be in such a world?" “

• Ben Kaube and Steve Smith on how generative AI is stripping publishers of their gatekeeping role by bypassing journal platforms altogether. I agree that authentic human engagement via a community is going to be the way forward, and I don't think scientific communities & their publications are going anywhere… and yet... Featured Notebooks in NotebookLM (mentioned above)… no reason why academic publishers couldn’t build a version of NotebookLM, but it would only work if it were a cross-publisher solution…

• Tony Alves’s sweeping and detailed survey of how AI is transforming every crevice of scholarly publishing, compiled from notes across six major 2025 industry conferences, is worth a read to understand where we currently are. Themes include AI's integration into peer review and editorial processes, ethical quandaries, fraud detection, global inequality in AI access, and the tension between efficiency and integrity.

• George Walkley’s excellent slides from the PLS Conference - AI: Threats and Opportunities for Publishers also give a broader summary of where we are and where we’re going. I made an incredibly last-minute decision to attend this meeting and thoroughly enjoyed it - really good conversations and thinking about how publishers respond to the threats and opportunities AI brings.

• Controversially, Law360 has mandated that all editorial content pass through an AI-driven bias detection tool, a policy introduced following internal accusations of political partiality. The system, developed in-house by LexisNexis, identifies language that may appear to lack neutrality, apparently prompting revisions even to factual or quoted material.

• 🎥 Inside Reuters’ AI playbook: Why Jane Barrett says journalism can’t afford to sit this one out

📚 Longer reads

• A new study from the Max‑Planck Institute shows that ChatGPT has begun influencing human speech. Analysis of over 740,000 hours of academic podcasts and YouTube lectures shows a significant surge in ChatGPT‑preferred vocabulary such as “delve,” “comprehend,” and “meticulous” since 2022.

• At last, a break from the breathless AI boosterism. HBR on The AI Revolution Won’t Happen Overnight “Yes, AI is powerful. Yes, it will change how we live and work. But the transformation will be slower, messier, and far less lucrative in the short term than the hype suggests.”

• Different datasets tell different stories in How Dimensions is illuminating scholarly publishing’s hidden diversity. Dimensions paints a hopeful picture of rising bibliodiversity whilst WoS sees power consolidating.

• 🎥 Live demo of Deep Background for Claude and ChatGPT o3 by Mike Caulfield. I’ve been playing around with the fact-checking part of this tool, and it’s good. You still need careful checking - e.g. tool can mark a fact as wrong, but a quick check of the original doc shows the tool has overstated a nuanced point.

• Why Metadata Enrichment Matters for the Public Knowledge Project Juan Pablo Alperin describes why PKP is deeply involved in efforts, such as the Collaborative Metadata initiative (COMET), to improve the completeness and accuracy of metadata.

• Eric J. Rubin and Kirsten Bibbins-Domingo defend medical publishing in a Washington Post Editorial.

• Maria Sukhareva explains why there are so many em dashes in AI-generated text and why we might be stuck with them. Let’s talk about em dashes

And finally…

This photo from Jr Kibs made me smile. I wouldn’t be surprised if someone had wanted to print this out for me back in the day. ‘Ask ChatGPT first’ might’ve spared a few sighs and raised eyebrows. Still, I miss asking people directly - and the conversations that followed.

📰 News

• 📚 David Worlock and Adrian Stanley explore how real, synthetic, and open datasets are shaping the future of AI. New systems are being built to support FAIR data practices, ethical licensing, and community-led stewardship so that AI development stays transparent, inclusive, and responsible.

• 📚 Whilst Anthropic purchased, shredded, scanned, and discarded millions of books to train Claude, OpenAI and Microsoft are backing Harvard’s Institutional Data Initiative, which is digitising nearly a million public-domain volumes to train AI in a gentler way.

• 🆔 This ORCID-commissioned white paper explores how persistent researcher identifiers are being used by funders, research communities, and commercial platforms.

• 😞 Sad news from the open infra world: Knowledge Futures begins winding down PubPub Legacy and Platform hosting after a key funder pulls out. Copim has published a new post explaining its decision to move from PubPub to WordPress. It’s a clear guide to the options they considered and the trade-offs involved.

• 💡 At SSP’s Shark Tank, five teams pitched solutions to the sector’s biggest headaches: from murky peer reviews to science no one reads. The winner? An AI-powered comms tool that transforms dense research into engaging public content

🚀Launches and new product features

• Cashmere helps publishers and authors maintain control over their intellectual property in an AI-driven world. It offers tools to manage rights and royalties, providing an ethical and structured approach to AI. Its focus on transparency and monetisation aims to future-proof traditional publishing against the rapid shifts in content consumption and automation.

• Developed by Laura Harvey, ScholarlyTrendSpotterGPT offers a thoughtful application of AI in the scholarly publishing sector. This custom GPT is designed to interpret announcements, news, and faint signals with sector-specific intelligence to help with strategic thinking, meeting preparation, or navigating industry developments. Also worth reading: Mark Carden’s comparison of the Researcher to Reader Conference with others in the field.

• Wax is a new open-source doc editor for teams. It has real-time editing, sharing, PDF generator, comments, Word import, and an AI designer, assistant, and knowledgebase - all AI tools are optional, off by default. (Blog post)

• Reviewer Three is a multi-agent peer review platform offering instant AI feedback on research papers which was built in 90 mins at the AI Engineer World’s Fair [H/T: @hoohar.bsky.social]

• The FAIA project offers a structured, transparent framework for disclosing AI contributions to digital content. Developed collaboratively by Liccium, Leiden University, and the GO FAIR Foundation, the initiative supports compliance with regulatory requirements, notably the EU AI Act, and encourages ethical, verifiable practices in content creation.

• JSTOR Seeklight is an AI tool purpose-built to support archivists, not replace them. It generates high-quality metadata via a multi-step process involving classification, prompts, and human oversight. It's been co-developed with practitioners and is part of JSTOR's broader digital stewardship ecosystem.

• Imagetwin’s AI-generated image detection is now out of beta. The upgraded tool no longer just flags suspect images; it also suggests which AI model, such as DALL·E, Firefly, or Stable Diffusion, most likely created them.

• World Brain Scholar has unveiled an enhancement to its editorial assistant Eliza, introducing automated scope checks and internal journal recommendations.

• Ian Mulvany has built a no-code citation converter using Claude.ai while still in bed. Users can paste references, pick a target style, and reformat with a click—when it works, that is. The “Failed to process citations” message reminds us that even magical tools have their Mondays. 😉

• A collaborative team of researchers, including Lisa DeBruine, Daniël Lakens, René Bekkers, Cristian Mesquida, Max Littel and Jakub Werner have developed Papercheck, an open-source, modular tool that screens manuscripts for adherence to best practices. From power analyses to data sharing, the software offers automated, pre-submission checks to help authors identify common issues early (Blog post).
  
  

🤖 AI & Publishing

• Keith Riegert’s “Getting Started with AI – USBS 2025” presentation is a practical yet slightly ominous tour of how AI is transforming every corner of book publishing. Everything from cover design and metadata to sales scripts, rights deals, and even the creation of entire books.

• A one-day Book Sprint held at the University of Vienna provided insights into the integration of AI within collaborative writing. Focused on Ludology, the study of games, the session employed tools such as Otter.ai and ChatGPT to assist in drafting, with mixed success. While transcripts proved a valuable record, participants critiqued the AI-generated summaries for their tonal inconsistencies and lack of nuance.

• Bryan Wilder explores the long-term, systemic risks of integrating LLMs into scientific peer review. Systemic risks go beyond accuracy; they could centralise judgment, incentivise conformity, and change how research is done, especially if authors start gaming the system or unconsciously tailoring work to please the machine.

• This fascinating and detailed case study explains how Ashlyn Wang used Streamlit and ChainForge to improve AI-generated newsletter headlines, focusing on factual accuracy and stylistic coherence. Manual and automated evaluations revealed measurable gains in consistency and editorial alignment 📈. (H/T: Nicholas Diakopoulos)

• The UK Government has just published a series of industrial strategies, including a Creative Industries Sector Plan. The plan includes a proposal for a new "Creative Content Exchange as a marketplace for selling, buying, licensing, and enabling permitted access to digitised cultural and creative assets.

• Bohyun Kim shared her interesting slides from her May ALA webinar talk, “AI: The Good, the Bad, and the Surprise,” on the role of AI in libraries.

• 📽️ George Walkley explains how AI is affecting copyright, creativity, and accessibility in publishing. It’s a measured and clear summary of where things stand today.

• 📌 I’ve put together a list of 30 People to follow at the intersection of AI and Academic Publishing. It’s part spotlight and part cheat sheet, and it could do with a little updating. Suggestions welcome!

📚 Longer reads

• 📖 Interesting to see how Anthropic’s new multi-agent research system works behind the scenes. A lead Claude agent spawns multiple subagents that run searches in parallel, compressing vast info into concise answers. It outperforms single-agent Claude by around 90 percent but uses fifteen times more tokens.

• 📽️ An excellent three minutes from Meredith Whittaker, President of Signal, as she cuts through the hype around AI agents. Is the tradeoff worth it? Handing over access to your browser, credit card, address book, Signal or Gmail, calendar, and more, just for the convenience of autonomous agents doing things for you?

• 📽️ I enjoyed the latest a16z podcast, Where We Are in the AI Cycle. I really hope “vibe writing” doesn't catch on. Highlights include the limits of autonomy, the shifting role of PMs, and the idea that instead of one "killer app," we may see new capabilities spread across platforms. The episode also touches on Google's future and the ongoing challenges of automation, especially with messy data and complex decisions.

And finally…

“Afterlife” is a wonderfully creative yet utterly disturbing example of AI storytelling. It's evocative, unsettling, and well-crafted. It pulls you into a world where bots and avatars feel eerily human and alive.

0

News

• 🧱 The 2025 State of Open Infrastructure report paints a rather bleak picture of open infrastructures left to fend for themselves due to volatile funding environments. See also Jean-Sébastien Caux's open letter to the scientific community reveals SciPost's financial struggles, with only 11% of benefiting institutions providing support.

• ⚠️ Emerald Publishing is confronting a surge in sophisticated phishing scams that impersonate their brand and staff to deceive authors into paying for fake publication services.

• 🌀 Laura Harvey shares reflections from 25 conversations about current shifts in scholarly communication: omni-crisis is the new norm, AI’s shifting to infrastructure, B2C is booming, and AI fear is real.

• 🎤 In his keynote at Publisherspeak UK 2025, Ian Mulvany highlighted the critical role publishers play in shaping future knowledge systems through deliberate infrastructure choices. He stressed that decisions made now, especially regarding identity management, will have lasting impacts on the integrity and accessibility of scholarly communication.

Launches:

• 🕵️‍♂️ James Heathers’ Medical Evidence Project blends white hat hacking tactics with journalistic urgency, taking a direct route to expose flaws in medical studies. Rather than wait on lengthy investigations, he’s hacking the process, literally and figuratively, by taking findings straight to the press.

• 📘DocuMark, courtesy of Enago and Trinka AI, promises to rescue university staff from the whack-a-mole game of AI detection. DocuMark enables students to verify their AI usage, gives educators clearer insights into student authorship, and integrates into learning management systems. [Press release]

• 🚨Preprint Watch aims to spot paradigm-shifting research early. By classifying preprints based on their potential to disrupt, it offers a sort of early warning system for those in publishing, funding, or policy-making.

• 🖼️After years of quietly crafting open-source tools for publishing and the wider scientific community, eLife is now putting it all on display. Their new website highlights how they’re helping build open tech for open science.

• 🕵️ Molecular Connections has launched a confidential integrity-investigation service named "2nd Opinion". The service offers speedy access to a team of Subject Matter Experts and integrity specialists from Molecular Connections.

• 💻 Adam Hyde’s Vibe Coding Workshops are live!
  

AI & AI longer reads:

• 📚 David Worlock in AI and Publishing: Death, Revolution – or Oblivion?, notes that while publishers are understandably focused on AI's role in content creation, the more significant change may be how readers themselves are beginning to use AI to access information directly. I’m no longer convinced that disintermediation is the right framing. In academic publishing, I think we’re probably entering a phase of increased intermediation that will be less visible and shaped by Big Tech.

• 🔍“Google Zero”, a delightfully ominous term for when Google sends no traffic to publisher websites, is the hot topic following Google’s rollout of AI Mode to search. 📉 For more, see Sara Guaglione from Digiday on What Google’s new AI Mode means for publishers, try The Rebooting Show or Pete Pachal on AI Mode, “...singular, concise answers may have the inadvertent effect of flattening knowledge diversity. Mainstream perspectives will get amplified, and niche or contrarian voices will have a tougher time standing out.”
  

• 🔗In his latest Substack piece, Mike Caulfield explores the differences in link hallucination and source comprehension across large language models. Using his trusty SIFT Toolbox, he finds that models that fabricate fewer links tend to grasp their sources better, because, shockingly, not making things up does help with accuracy. ✅ More seriously, it’s a timely reminder that we need more benchmarking of this kind for academic contexts, and a growing suite of AI-wrapper tools to vet content not just for correctness, but for how well it aligns with sources.
  

• 🧠Rosalyn Metz examines the Wiley-Perplexity partnership, highlighting its potential to enhance academic content accessibility through AI-driven search. 💭I can’t help thinking that the future for both publishers and librarians will be somewhat bleaker. I don't think papers will remain the primary access point for academic STM knowledge. Instead, it will likely be through Big Tech platforms managed via the university rather than the library, with content filtered in ways many of us find uncomfortable but will accept for the sake of convenience. On more optimistic days, projects like the National Library of Norway’s national LLMs suggest a future that could still be more open, diverse, and library-led.
  

• 💡Chris Duncan floats a radical idea: what if news publishers joined forces to buy Chrome in a bold bid to wrest back control of content distribution and monetisation? But will we still need web browsers? 🕸️ Scott Rosenberg suggests we might be witnessing the browser's last hurrah, with browsers and search increasingly bypassing traditional clicks and content in favour of summaries and chat interfaces. This transformation, led by companies like Google, OpenAI, and newer players like The Browser Company, threatens to unravel the economic and creative incentives that kept the web vibrant in the first place.
  

• 🧨Anthropic CEO Dario Amodei recently warned that AI could wipe out up to 50% of entry-level white-collar jobs, possibly pushing unemployment to 10–20% in the next five years, causing considerable comment. 🧮 More hopeful takes come from Nathaniel Whittemore, reading and reflecting on Tim O’Reilly’s AI First Puts Humans First, and Jules White’s question: “Will companies use AI to help everyone do more—or just cut jobs and keep output the same?” White makes the case for the first option: “Empower your people. Expand your vision. Let AI scale your ambition, not shrink your workforce.” 💼 I’d like to believe that, but I think scholarly publishing will go the other way. The focus on efficiency and protecting profit margins makes a “do more with less” mindset hard to shake.
  

• 🧠 I learn so much from Aaron Tay and really appreciate the time he takes to put together his posts. Here are three great reads:

  • A Librarian’s Guide to AI in Academic Search Tools. A really good, but long, read that deserves your time.

  • How should academic retrieval augmented generation (RAG) systems handle retracted works?

  • Generative AI in Practice: EBSCO’s AI Insights by Richard Bridgen (found via Aaron's Bluesky feed). Interesting to see how this was tested and the conclusion: “Given the issues the testers raised to do with missing detail, inconsistency, the potential for bias and lack of customisation of the feature, it was decided not to switch it on until such time as these issues have been addressed.”

• 📺 Benedict Evans’s latest presentation, “AI eats the world”, is a sharp, accessible summary of where we are with AI today. Not publishing-specific, but a great contextual overview.

  

• 📋 In case you fancied a light 167-item browse, I’ve refreshed the AI-Powered Publishing Use Cases list.

And finally:

🦀 The Shawshank Crustacean. A lovely crash blossom: "The convicted killer crab walked between two walls and pushed through newly installed razor wire..."

• Let’s start with some positive vibes: SciFlow is offering free accessibility checks for PDFs ahead of the European Accessibility Act, which comes into effect next month. (H/T: Lou Peck)

• ScholarOne, the ageing engine behind over 20% [❗] of the world’s peer review, has long been a drag on progress in academic publishing. Silverchair describes its plans to overhaul this clunky colossus and reconnect with customers.

• A COAR survey finds that over 90% of open repositories are regularly disrupted by overly aggressive AI bots, leading some to restrict machine access altogether. These measures have unintentionally impacted legitimate services, prompting COAR to establish a Task Group to develop more measured mitigation strategies.

• Wiley, via Josh Jarret, has introduced an AI agent on AWS, marking a move towards AI-driven access to academic content. It’s an intriguing development for those betting on AI as the future of research discovery. Still, given that publishers have been licensing content to biomedical AI firms for years, it’s perhaps less groundbreaking than it first sounds. For a more sceptical take, see Kent Anderson’s commentary.

• Wiley, also via Josh Jarret, has also partnered with Perplexity AI to make its content in their generative search and become Perplexity's first education partner. [It can’t be easy living up to a title like SVP and GM of AI Growth; every week must demand a new announcement!]

• To mark its 25th anniversary, Crossref has launched the inaugural Metadata Awards 🏆, recognising six organisations for outstanding contributions to scholarly metadata. The awards highlight the importance of high-quality metadata in enhancing the discoverability and integrity of research.

• Finally, a quick recap of the AI Publishing Collective/Cambridge AI meetup, join the LinkedIn group to get updates on the next meeting.

🚀 Launches/new to me

• ReviewerZero is a new platform designed to catch research integrity issues before publication. It covers things like authorship verification, statistical checks, image duplication, and citation problems.

• AI-Powered Publishing: Use Cases is a work-in-progress list of how academic publishers are using AI. It's currently light on Generative AI examples, so feel free to suggest additions.

🔎 Research Assistants

• I came across three new tools this week Corpora.ai, Ai2 Paper Finder, and FutureHouse’s AI Agents for Scientific Discovery. Aaron Tay offers an in-depth look at the last two of these tools in his latest post.

• Aaron is also offering a "Master Class" on May 29, 2025, from 4 PM to 5:30 PM (SGT), aimed at librarians exploring how AI is reshaping academic search.

🤖 AI

• More data is emerging on the impact of Google’s AI Overviews. “The study analyzed 400 AIO (AI Overview) interactions from 70 users. Click-through rates drop by up to 66 percent on desktop, and nearly 50 percent on mobile. Seventy percent of users only read the top third of the AI answer box. Younger users trust AI, older ones want sources.”

  

  I suspect we’re watching the slow handover of academic information control from publishers to platform-driven intermediaries, and barely anyone’s blinking. # Ben Kaube is thinking along similar lines: The Stack Overflow Effect: Are Scholarly Publishers Sleepwalking into an AI Disintermediation Crisis?

• Aaron Tay shared a thoughtful Bluesky thread outlining key librarian concerns about generative AI, which include hallucinations, user dependency, environmental impact, and copyright, and how these are being addressed or could be approached differently.

• 😮‍💨 Sigh… Morelle’s Statement on Abrupt Firing of Shira Perlmutter, Register of Copyrights. As Seven Fund comments: “Publishers need to tune up their voices in defence of their values… Publishers still hope that strategic silence is their best option – I am sure it is not.” Agree. Publishers are, at times, a touch too fond of staying quiet.

• Nature’s survey of 5,000 researchers uncovers a distinctly mixed view on AI in scientific writing: widely accepted for editing, met with some hesitation for drafting, and rarely used for peer review. The charts add colour, showing generational splits and a landscape of opinion that ranges from curious optimism to raised eyebrows. Disclosure, it seems, is still something of a grey area.

  

• ⚖️Peter Csathy’s AI Litigation Tracker (updated 11 May) is an oddly compelling who's suing whom of the generative AI world, mostly from the media and entertainment sector.

📚 Longer reads:

• Academics: Little more than intellectual orphans in AI era? Interesting piece on how AI is evolving from a helpful academic tool to a central force in knowledge creation, peer review, and publishing. Do academics risk becoming intellectual orphans, detached from their work & the institutions meant to support them?

• Microsoft’s AI in Education report throws up a quietly intriguing stat: 47% of education leaders are using AI daily, while only 7% of students are. For all the talk of digital natives, it seems institutional decision-makers are currently the ones experimenting most with the tech:

  

• Humanities Commons has given its backup systems a thorough workout and, to everyone's relief, they passed with flying colours.

• AI in Media: The Infinity Paradigm by Ezra Eeman offers a comprehensive overview of where we currently stand with AI in media, and where the road ahead might lead. 🔮

• An MIT PhD student briefly grabbed the spotlight with claims that AI was fuelling breakthroughs in materials science - until it emerged the data had been fabricated.

A Pioneer in Publishing and AI: Remembering Einar Fredriksson

I was really saddened to hear about the death of Einar Fredriksson. He was full of ideas and always generous with his time. Einar was one of the early pioneers in thinking about how AI could be used in publishing, and I really valued his big-picture thinking. He believed strongly in open science and FAIR data, as his most recent work with LIFES and the GO FAIR Foundation showed. I’ll miss his wisdom, curiosity, and kindness.

🔚 And finally…

Seven books, one reader, zero screens: the 1700s' answer to information overload.

• 🔢 The Wikimedia Foundation has teamed up with Kaggle to launch streamlined Wikipedia datasets for AI training, hoping to ease the load on Wikipedia’s servers. Curious to see if it’ll catch on, Wikipedia might just have enough clout to make it work. It’s been nearly 20 years since platforms like HighWire tried offering similar services to publishers, only to be roundly ignored. The Data Licensing Alliance gave it another go in 2021 with much the same result. Perhaps the third time’s the charm?

• 🔧 Open knowledge doesn’t preserve itself. Rosalyn Metz on The Fight for Open Infrastructure Starts Now (Actually, It Started Yesterday)

• 🍪 “After more than half a decade of delayed timelines, government oversight and industry infighting, Google has fulfilled many expectations: a U-turn on third-party cookies in Chrome.” Digiday, on how the ad-supported media ecosystem has responded to what, in hindsight, has been a costly distraction.

Launches

• 📈 The Data Catalyst newsletter, Thoughts and news on data quality, semantic interoperability, AI, FAIR data, and more (Healthcare and Life Sciences focus) by Andrea Splendiani

• 🔍 SCiNiTO (this has been around for a while, but it’s new to me) offers full-text access to Open Access sources and institutional holdings, plus a few AI bells and whistles: an AI Chat for research assistance, an AI Reviewer Agent for manuscript evaluation, and a Journal Recommendation system to help you find the best place to publish.

• 📚 Reedy (and here I’m even slower off the mark!) is an AI-driven platform designed to simplify and optimise metadata management for books.

AI

• 👉 London-based? Don’t forget our next in-person AI in Publishing Collective meetups on 29 April and 14 May.

• 📊 Leslie McIntosh on why trusted data matters. “Large, linked datasets fuel AI tools, and that’s exactly why trusted data matters. Without it, we can't expect trustworthy outputs.”

• 📹 Michael Upshall has a nice write-up of the Fiesole Retreat 2025. I love this anecdote: "Paris O’Donnell, from the library of I Tatti, described a clever way of checking for missing or incorrectly shelved books in the library: simply videoing the sequence of books on the shelf and comparing it, using vibe-coded AI, to the library catalogue."

• 🤖 Florent Daudens reports about what might happen to news publishing if people stop actively embracing the web, paraphrasing:

  Picture this: AI agents will be able to browse the web for us faster and better than us, without seeing the ads, We're producing more content than ever, AI will intermediate content for us, be it with our own tailored agents or without us realizing it (à la TikTok). Add these three trends: Models are being commoditized, so tech giants are racing past basic models to build products that will capture user attention, Smart devices will embed smaller but more capable models, giving users unprecedented control over how they want to get their news and content, Our culture is rapidly shifting from written to oral communication, This could mean the death of the web as we know it, as well as web advertising, just like digital almost killed print.

  If people stop using the web, a big if, I think the impact on academic publishing will likely be slower and more nuanced than in news media, but I do think AI-intermediated content will present a real challenge, especially as aggregation risks stripping away important trust indicators and context (see Leslie McIntosh’s article above). The point about our culture shifting from written to oral communication is particularly interesting to think about.

• 🤩 AI tools are overhyping research results, according to a new study, with the newer models being the worst offenders. Not exactly surprising; it takes skill to summarise findings without stripping out all the nuance (I used to do this in my first publishing job and never got it quite right… ).

ChatGPT’s distillation of Springer Nature’s Annual Report

Annual reports are curious things — they tell the story a company wants you to hear, which isn’t always quite the same as what’s actually going on. Combine one of these reports with ChatGPT’s boundless enthusiasm, and you get the following Technology & AI Strategy Deep Dive based on Springer Nature’s FY2024 results:

🔍 Executive Summary

Springer Nature (SN) is rapidly evolving from a tech-enabled publisher to a tech-native operator. Their 2024 results showcase significant progress in applying AI and automation at scale across editorial, production, and product functions. Unlike many legacy peers still piloting GenAI, SN has embedded these capabilities across key revenue and operational workflows.

SN’s public commitment to “90+ AI pilots” masks a more sophisticated maturity: they are now internalizing tooling and platform components that structurally reduce cost, compress cycle times, and subtly redirect user behavior (authors/editors/librarians). Competitor publishers should monitor closely not just what SN builds, but how it operationalizes and integrates it.

---

📈 Key Technology Takeaways

1. AI at Scale: From Pilots to Process

• 90+ AI pilots run across all business units, indicating strategic experimentation beyond Research.

• Tools such as Snapp (submission-to-accept platform) now handle >1 million submissions annually—a custom-built backbone for their editorial operations.

• AI is no longer just R&D—it is production-grade and business-critical.

Implication: SN is outpacing competitors in turning GenAI into operational savings and workflow automation. Expect their cost base to become structurally leaner over time.

---

2. Tool Ecosystem: Modular & Use-Case Specific

SN has deployed a suite of targeted AI tools, each designed for a key pain point in the publishing lifecycle:

Implication: Springer’s tooling is focused on automation and user retention—each tool solves a workflow problem while also reinforcing author/editor loyalty within their ecosystem.

---

3. AI as a Product: Nature AI Assistant

• SN piloted its Nature-branded AI assistant with ~600 researchers in 2024.

• 81% said it saved them time in reading/writing tasks.

• Open beta launch is expected in Q2 2025.

Implication: This shifts SN’s AI from back-office tooling to revenue-generating user experiences. Positioning AI as a branded companion (not just functionality) opens doors to subscription, licensing, or institutional bundling models.

---

4. Data + Design: Subtle Platform Power Moves

• eBook usage up 35% YoY—driven by data-informed library offerings like “Access & Select.”

• SN uses interaction data to refine product design, automate recommendations, and drive purchasing decisions.

• Internal rollout of an AI Academy trained 2,000+ employees.

Implication: SN’s tech is not only externalized—it’s internalized. Their workforce is being upskilled to normalize AI in everyday decisions, likely accelerating future tooling uptake.

---

5. Digital Maturity by Segment

Implication: Research is SN’s innovation engine. EdTech is a secondary frontier. Health remains a laggard, with limited signs of disruptive tech.

---

🧭 Strategic Watchpoints for Competitors

1. Cost Curve Advantage: SN’s 60% production cost savings via ACDCx will pressure competitors still using third-party typesetting vendors.

2. Sticky Ecosystem Design: Tools like T-Rex and Journal Finder reinforce author loyalty and increase submission velocity—a long-term market share play.

3. From Service to Product: Nature AI Assistant may evolve into a sellable subscription or institutional tool, unlocking new direct researcher revenues.

4. Employee Enablement: The internal AI Academy suggests a deeper change management program—expect faster rollout of future tools than peers.

5. Branded AI Arms Race: Nature branding gives SN an edge in trust and adoption. Expect other publishers to rush to brand their own AI companions.

---

🧠 Analyst Commentary

“Their innovation narrative is unusually coherent—from TAs and OA to automation and AI assistants. Most publishers can’t string that together yet.”

Here’s how ChatGPT sizes up the technology strategies at Springer Nature, Elsevier, Wiley, and Taylor & Francis. I think there’s some truth to the comparison, though Wiley and Taylor & Francis might see things a little differently!

Longer reads:

• 🔖 WIRED’s article explores how arXiv has quietly transformed global science by enabling researchers to share preprints quickly and freely. arXiv’s impact has been immense despite a rather retro façade.

• 🔖 Is it time to trade in the science journal? Once the reliable Volvo of research, are they becoming more Morris Minor-like, charming but entirely unsuited to the motorway of modern science? Worlock delivers ten firm nudges toward the exit, citing slow speeds, spiralling costs, and a distinct allergy to AI.

• 🔖 Money, Trust and Tools, How Two New Projects Aim to Strengthen Diamond Open Access by Elena Di Rosa (H/T: ‪Samuel Moore‬ on Bluesky)

• 🔖 In 2025, Gen AI seems intent on helping us feel a bit more together - or at least a bit less scattered. Top uses now include therapy, life admin, and finding purpose, while text editing and content polishing have quietly shuffled down the rankings. 📈 Interesting related stat: character.ai reportedly handles 20,000 queries per second—around a fifth of Google’s search volume.

• 🔖 Academic Databases and the Art of the Overcharge. “Clarivate, Elsevier, and the American Chemical Society are comfortable pursuing a strategy of pricing discrimination. But there is no reason libraries should allow this approach.” Alternatives like The Lens also exist.

And finally…

😂🎯 Every AI copyright argument you’ve ever heard in bingo form. Genius from Graham Lovelace:

News:

• The new STM Trends 2029 graphic explores the convergence of AI, societal fragmentation, and the erosion of trust in research. It advocates for new coalitions to rebuild integrity in scholarly communications.

  

• 📈Interesting stat: Springer Nature “... recently conducted a pilot study simply requiring authors to explain why any unshared data hadn’t been deposited in a public repository before final acceptance. This request alone raised data-sharing compliance from 51% to 87% in participating journals."

• 🔍 Monash Health Library has rated AI search tools "….we determined that Consensus, Evidence Hunt, Lens.org, Semantic Scholar were the most useful tools, having a ranking of 9 out of 10. Elicit, Litmaps, OpenAlex & Scinapse closely followed with 8 out of 10". (See also Aaron Tay’s comments on Bluesky)

  

• 🦋 Bluesky surpasses X as the preferred platform for academic research sharing, according to an analysis by Digital Science.

• 🎓 The Verge believes that OpenAI and Anthropic are racing to dominate the higher education space, each launching new tools tailored for university students. While OpenAI offers perks like free ChatGPT Plus access during finals, Anthropic’s Claude leans into Socratic learning to deepen student engagement.

🚀Launches

• 🕵️‍♂️ Oversight by Clear Skies. The first comprehensive index of journal integrity. Using machine learning and network analysis, Oversight analyses data from 60M+ papers (since 2016) to detect signs of organised research misconduct. Each journal is assigned a Composite Risk Score, enabling publishers to identify potential integrity risks within their portfolios. By going beyond traditional metrics like publication volume and citation counts, Oversight aims to offer a more nuanced and data-driven perspective on journal quality:

• 🔢Frontiers introduces FAIR² Data Management, an AI-assisted service designed to improve the structure and accessibility of research data.

• 🔍 Silverchair & OUP launch the AI Discovery Assistant on Oxford Academic, boosting research with smarter search and AI summaries.

• 🤖 Clarivate announces "...expansion of its Academic AI Platform, with the aim of helping institutions harness “agentic AI” to accelerate productivity, save time for researchers, students and staff, and engage users..."

• 👉 Unriddle has changed its name to Anara and will be focusing on researchers: “Anara helps you understand, organize and write scientific documents with AI. ”

• 🔍 Scite Assistant has launched a powerful new AI model focused on advanced reasoning, significantly improving scientific response accuracy, problem-solving, and information synthesis. It’s currently available for Enterprise users.

• 🕵️‍♂️ Springer Nature has launched an AI tool designed to detect irrelevant references

• 📄 JSTOR has launched Digital Stewardship Services, a comprehensive platform designed to assist libraries and archives in managing, preserving, and sharing their digital collections. The service integrates AI tools for metadata generation, aiming to streamline workflows and enhance discoverability.

• 🤖 Zendy's new revenue-sharing model leverages its AI assistant, ZAIA, to compensate academic publishers when their paywalled content is cited.

AI-related reads:

• 🧠 Hum sees itself as the brain & nervous system of modern publishing. Unifying data, harnessing AI, and offering tools like Alchemist Taxonomy, Insight, Search, & Review to streamline the whole affair.

• 🔖New IBM report reveals that the journey to an AI-powered content supply chain is hitting real-world snags. 🚧 Early adopters are encountering unexpected complexities - from integration snags to scaling issues.

• 🤖 Il Foglio's bold experiment with a fully AI-generated newspaper insert is facing challenges. Think misspellings, muddled facts, and more meh than marvelous.

• 📚 Copyrighted books in OpenAI’s training data? Spoiler: Yup (probably). Method = clever.

• 🧘🏿 In 2025, Gen AI seems less focused on dazzling us with code and more intent on helping us feel a bit more together. Top uses now include therapy, life admin, and finding purpose, while text editing and content polishing have quietly shuffled down the rankings: https://hbr.org/2025/04/how-people-are-really-using-gen-ai-in-2025

Longer reads:

• 🔖Michel Upshall explores institutional repositories. With machine learning and automation creeping in, are these overlooked engines of open access on the cusp of a quiet revolution?

• 🔖Jeffrey Robens emphasizes that writing is an integral part of the thinking process for researchers, warning that relying on AI like ChatGPT for writing may hinder idea development and accuracy. I absolutely agree that for more discursive, conceptual work, writing is integral to the thinking process. But for the more formulaic end of research – the "I did X and Y, and got result Z" type papers, I wonder if we should be adapting the structure of papers and automating more? Do we really need every researcher to write another slightly reworded paragraph on why 'wireless sensor networks' are vital just to clear the plagiarism filter?

• 🔖Haseeb Irfanullah argues that the traditional peer review system has become impersonal and inefficient, plagued by reviewer fatigue and procedural rigidity. He advocates for a transformative overhaul to restore its collaborative essence and adapt to modern challenges.​ See also 🎧 episode 190 of Everything Hertz, titled "What happens when you pay reviewers?", features hosts Dan Quintana and James Heathers discussing two studies that assess the impact of compensating peer reviewers on the efficiency and quality of academic peer reviews.

And finally…

📚 ☺️ ShelfImage is a lot of fun: shelfimage.co.uk. Snap a photo of your bookshelf and let this clever little tool gently tease out the quirks and curiosities of your reading taste. According to my latest book pile, ShelfImage says:

Your Bookshelf Personality: The Visionary Entrepreneur

This collection screams innovation and self-improvement with a dash of creative culinary experimentation. You're someone who thrives on new ideas and isn't afraid to explore the unknown, whether it's in the realm of AI or sourdough starters.

Some Of The Books We Spotted

• Surviving and Thriving in the Age of AI

• Ferment from Scratch

• Body of Work

• The Start-Up of You

• AI in Business

Your Personality Traits

• Innovative

• Curious

• Ambitious

• Adaptable

🚀 Hum and GroundedAI launched Alchemist Review, an AI-Driven Peer Review Solution. Now being tested by select journals, it aims to tackle the age-old publishing woes of reviewer fatigue and snail-paced turnaround times. See also: Ann Michael introduces the Peer Review Assistant (PRA). [More to come on this in a future issue!]

📉 The launch of Google’s AI-powered search summaries has led to a 19% drop in click-through to academic reference services according to OUP product strategy director John Campbell. "Discussing what he called the advent of “zero-click journeys” for academic publishers, Campbell explained how half of all Google keyword searches likely to surface information within OUP’s own platform, Oxford Academic, now appeared with an AI-generated description next to them." [H/T: MediaMorph]

🚀 The Disentis Roadmap, an ambitious 10-year plan to unlock research data currently buried in scientific papers. By 2035, the initiative aims to create a ‘Libroscope’—a pioneering system that will make biodiversity knowledge freely accessible, AI-ready, and fit for purpose in science, policy, and conservation.

💰Invest in Open Infrastructure (IOI) has launched its call for proposals for the IOI Fund for Network Adoption. With support from major philanthropic organisations, the fund aims to boost open data and content infrastructure across Africa, Latin America, and North America.

🗂️ openRxiv is the newly launched nonprofit home for bioRxiv and medRxiv, aiming to ensure that life and health science preprints remain freely accessible and researcher-led. Science article.

🚀 CCC will launch a new AI Systems Training License in late 2025 to help organisations legally use third-party content in AI training and output.

🛠️ EasyJournal, an experimental open-source prototype journal platform developed by Adam Hyde, is looking for early testers. More about the experiment.

Peer review

🤖 Japanese startup Sakana claims its AI tool, The AI Scientist-v2, successfully generated a scientific paper that passed peer review at an ICLR workshop. However, the paper, which was later withdrawn, contained “embarrassing” citation errors, such as wrongly attributing a method to a 2016 paper instead of the original 1997 source. In a candid blog post, the company also admitted that none of its AI-generated studies met the internal standards required for submission to ICLR’s main conference track. [TechCrunch story]

🔥I think YesNoError sounds like an interesting experiment but I also like Carl T. Bergstrom’s Bluesky post about the writeup in Nature, “Is everyone huffing paint? Crypto guy claims to have built an LLM-based tool to detect errors in research papers; funded using its own cryptocurrency; will let coin holders choose what papers to go after; it's unvetted and a total black box—and Nature reports it as if it's a new protein structure.”

🐶 Neil Blair Christensen is shamelessly resorting to adorable dog pics to nudge folks in peer review and editorial circles to take a 1-minute survey on publication integrity. Survey questions I Dog pic

AI

📰 A controlled study by Nick Hagar investigated whether, based on someone’s reading history, an LLM could adapt the content for the newsletter in an engaging way. Results, readers preferred AI-generated newsletter headlines 62% of the time.

👎No great surprises here, but useful to have the data: a Tow Center study shows that generative AI search tools often struggle with citing news accurately. Even platforms with licensing deals don’t show much improvement when it comes to giving credit where it’s due.

🔧A new report from FT Strategies in partnership with the Google News Initiative, guides publishers through evaluating cost, feasibility, risk, and strategic alignment when facing the “Build or Buy” dilemma. It’s aimed at News publishers but the helpful concluding questions to ask yourself as you consider AI deployment apply to most companies.

🧪 A quick project by Ian Mulvaney using AI tools like NotebookLM, DeepSeek, and Claude to highlight M&A activity in scholarly publishing from 2021–2025. Drawing on ~30 sources (including Clarke & Esposito's newsletter), it efficiently produced a visual snapshot of industry consolidation.

📈 New Adobe research shows a 1,200% jump in AI search referrals to retail sites in 2024–25. Shoppers are turning to tools like Perplexity, ChatGPT, and Google AI for product research (55%) and shopping (39%). Bain Generative AI Consumer Survey says that about 80% of search users rely on AI summaries at least 40% of the time.

Longer reads

🔖AI isn’t making intelligence ubiquitous—it’s commoditising expertise. Tim O’Reilly argues that true intelligence lies in creativity, values & taste: the ability to see what others miss and make meaningful choices.

🔖I finally read Geoffrey Bilder’s On AI from 2023. His call to stop calling it "AI" may have already been lost! He argues metadata & context—not flawed AI detection—are key to research integrity. In 2025, we’re seeing more context-based approaches coming from start-ups and other organizations, but will any of them be community-driven?

🔖 Molly White’s “Wait, not like that”: Free and open access in the age of generative AI explores how generative AI challenges the free and open web—and how restrictive responses could do more harm than good. Could solutions like fair compensation, ethical AI use, and sustainable frameworks protect open knowledge? Maybe… Hopefully…?

🔖 Bergstrom & West deliver a succinct and measured 'lesson' on AI in science - covering protein design breakthroughs, Sakana AI's $15 AI-written papers, and peer review by LLM. A thoughtful take on where AI helps, and where it’s just hype.

‪

Events

📅 AI Publishing Collective meetup in London, keep 29 April and 14 May free in your diary! Join the group on LinkedIn for updates.

📅 BCS Knowledge Discovery & Data Mining series on May 14 (6-7:30 PM BST) has 2 talks by Anna Fensel on knowledge graphs, FAIR principles, and generative AI in agri-food, and Anelia Kurteva on responsible AI through data governance that might be of interest. More details & Zoom link (website is http and flags up a security warning, zoom link for the event with no registration needed is https://ukkdd.org.uk/2025/?section=registration).

📅 AI and the Future of News 2025, 26th March 2025, A one-day conference with speakers from journalism and academia. Free and online.

And finally…

❤️ I absolutely love this from the University of York Library on Bluesky: “Not that we in the library wish to encourage profanity, but if you want to get rid of these AI-search summaries (which are often absolute nonsense), just swear in your search terms.”

🆕 New, newsworthy & new to me

• Libraro, a blockchain-based platform that gives writers the chance to bring stories to life, is crowdfunding 💷 [Pitch on Crowdcube, requires login I LinkedIn post]

• SciScore and bioRxiv Launch New Key Resource Tables for Preprints

• Turn ChatGPT into Grammarly with EditGPT

• SprinerNature was advertising for a PM to work on a ‘loyalty SaaS solution’ "...driving long-term publishing success... by selecting and integrating a loyalty SaaS solution to recognize and celebrate diverse research contributions while offering valuable benefits to researchers". The Times They Are A-Changin'!
  

✍️ Peer Review & Peer Review platforms

It feels as if the sleepy and somewhat moribund* world of peer review platforms has suddenly come to life:

• Molecular Connections acquired Morressier and Editorial Office Ltd. More about Morressier acquisition in Scholarly Kitchen analysis by Roger C. Schonfeld

• River Valley Technologies Releases Version 3 of ReView which now includes Integrated Research Integrity checks and annotated Peer Review

• ScholarOne has published their 2025 Product Vision focusing on research integrity, rethinking the Author, Reviewer, and Editor Experience, modernization, and community-focus

• Reviewer Credits has a new lease on life under Sven Fund’s and a newish podcast
  

🤖 AI

• Avi Staiman, in Science Editor, calls for more nuanced AI policies in publishing, advocating for a risk-based framework: AI use and declaration: When Declarations Just Don’t Cut it: Building a Risk-Based Framework for AI Guidelines in Publishing

• Kent Anderson’s Deep Research = An Ignoramus is a reminder of how wrong GenAI can be when data is scarce—though it often delivers “good enough” results for most use cases.

• Adam Hyde’s Documenting My Assumptions on AI and Schol Comms is a thoughtful look ahead, predicting rapid AI coding advances, collaborative user agents (CUA), and digital worker evolution. I agree with Adam’s thinking, one area I see coming is a shift from text to voice interactions with AI. I recently saw a demo where someone created a full presentation using ChatGPT—research, summarising text for slides, and generating visuals—without touching a keyboard.

🎯 Interesting opportunities

• Peter Suber is seeking expressions of interest from nonprofit orgs to take over the Open Access Tracking Project and the open-source software on which it runs.

• Deanta has just launched its annual Trends in Academic Publishing Survey and wants your opinion

• Our first AI, Publishing, and EdTech meetup went well, join our LinkedIn group at https://www.linkedin.com/groups/13180046/ for updates about the next event. Michael Upshall has a good write-up of Phil Gooch’s great talk Nine Rules for a Start-up. If you would like to speak or can help with a venue please get in touch!

  
  

📖 Mind-expanding longer reads, watches, and listens:

• 🎙️ The latest episode of Midnight at The Casablanca with Paul Peters has a section about the development of Phenom - which reminded me how dynamic the open-source space was from 2018 to 2020.

• 💬 Mike Taylor, Head of Data Insights at Digital Science, shared A first look at Altmetric insights from Bluesky, in mid-Jan, “the most shared piece of research is “A Place of Joy”, an article published in Nature about the growth of the research community on Bluesky. The oldest article is a 1684 account of an earthquake near the ancient city of Constantinople, while the most viral papers (those that have the highest ratio of reposts to original posts) are those that discuss issues relating to DEI in academia, and the use of generative AI.”

• 💬 Tom Chatfield in Thinking out loud onscreen argues that “the need to reach out to one another through felt and crafted words will endure even while who “we” are will become a question inseparable from the technologies that ceaselessly mediate the world for us”

• 💬 Luc Wiesman’s Google's Algorithm Is Turning Original Journalism Into Zero Sum Game for Small Publishers is a stark read. As Google’s organic traffic plummets—down 90% for some—Australian editorial teams from Carsales to LadBible are gutted, threatening niche publishers' survival.

• 💬 Are We Making the Most of Digital Forms of Scholarly Communication? in Katina Magazine notes that even back in 1922, W. J. Crozier declared that “the journal method of publication had become antiquated”.

• 📺 The JournalismAI Festival was a fantastic and incredibly diverse event, try starting with Journalism and AI - Future Trends

• 💬 Kim Eggleton wishes for “submission system providers, please develop your APIs and collaborate with the new integrity tools entering the market; and emerging start-ups, consider integrating your innovations with existing systems – it will help your products reach users faster and allow us collectively to build a more efficient and trustworthy publishing ecosystem”. Highwire, Clear Skies, and DataSeer have done just that. [Press release]

🖼️ And finally…

I loved the ⚡Electric Dreams⚡ exhibition at Tate Modern. It's subtitled Art and Technology Before the Internet but much of it still feels modern, radical, and experimental. It's also a lot of fun, capturing that wonderful 1960s optimistic vibe about how tech can positively shape the future. Well worth a visit!

* Ok - this comment really only applies to the big platforms, I know there’s been a ton of exciting happening with the smaller platforms 🙂