This is a series about Wikipedia's anonymous editorial power — and what happens when AI turns contested human judgments into facts.

Don't miss Part I, Part II, Part III, and Part IV.

Subscribe

They don’t make it easy but I promise I’m really trying to end this series on a somewhat positive note. The last four installments have largely been a long-form complaint about all the things Wikipedia’s getting wrong, without sufficiently acknowledging one of its better features that’s been there all along. The people making the decisions may be anonymous, but at least the decision-making itself is public.

It’s good for obvious reasons —when a random stranger adds a contested sentence about a person, that sentence sits inside an article that comes with an edit history and a talk page, sometimes a footnote, or if you’re lucky maybe even a warning box informing you that the neutrality of this section is disputed and the content is outdated. So although the system is generally unaccountable, it remains fairly legible, and imperfect as it might be, you still have the option to watch the process unfold and judge for yourself.

It’s fair to say that Wikipedia was never designed for a lazy reader with dormant critical thinking skills, and it certainly was never meant to be blindly fed into LLMs after being stripped of all context. And yet that’s exactly what’s happening now — AI models inherit the sentence while the discussion trail is left on the cutting-room floor.

Wikipedia is Internet’s everything bagel from Apollo, minus the lunchtime line. When you think about what an AI company needs — enormous body of text, consistent structure, uniform tone, dense cross-referencing, along with full coverage of Byzantine coinage, Belarusian politics, entirety of The Simpsons, list of every man who ever got to play forty-five minutes of professional football, written in English and maintained around the clock by people who will never send you a bill, all while being licensed so freely that nobody even has to ask. Almost nowhere else online offers that combination. AI companies have scraped plenty of material without asking, of course. The difference is what happens afterward. Books and news articles come with copyright holders who can sue, and licensing departments that take forever to answer and then want to be paid. Then, years of litigation over where fair use begins and ends. Wikipedia comes with permission already attached, one checkbox, all-inclusive.

And it didn’t get here by winning some local spelling bee or because the planets happened to align in its favor. Search engines helped cement Wikipedia’s authority by placing it at or near the top of results, year after year.

Researchers have found Wikipedia links appearing on between 67 and 84% of desktop search-result pages for common and trending queries across the engines they studied, with earlier results on Google putting it above 80% for some major categories, and the reason provided was the push to “combat misinformation”. It became a compounding engine - search sent readers, some of whom became editors, who then expanded the content library, which in turn ranked for more searches, rinse and repeat. Until it stopped.

The Knowledge Graph was introduced in May 2012. It displayed an extracted answer in a panel at the top of the search results, with the launch post openly naming its sources—Freebase, Wikipedia, and the CIA World Factbook. By January 2014 The Register was reporting that English Wikipedia's pageviews had fallen 21% over the previous twelve months (12% once mobile was counted), with German, Spanish and Japanese down even more, and it pointed at the panels. A Wikimedia spokesperson said at the time that they had not noticed a significant drop in search referrals, and noted a continuing dialogue with the Google staff working on the panel. Six years later the scholar Dariusz Jemielniak made the obvious point — most people receiving those answers had no idea they were reading Wikipedia, which meant fewer donations and fewer volunteers.

AI Overviews are hardly a conceptual break from any of this, they’re merely an upgraded version of the same idea. Google helped build Wikipedia, then learned to extract the answer on its own, giving the audience quick tailor-made summary instead of a lengthy boring article. Throughout all this, our good encyclopedia remained free to read, free to edit, free to license, and free to take. Its work now sits as a quiet foundation beneath search summaries, voice assistants, answer engines, social media descriptions and AI-generated slurp distributed across the rest of the internet.

Nothing’s more expensive than a free lunch, and it’s always fun to learn the hard way.

“And now the court [meaning every Wikipedia admin/editor], is freelancing for your favorite chatbot.” I ended Part III with this line half-jokingly, only to discover, after doing some more research, how accurate this actually was. So I obviously had to pause and congratulate myself, especially since I haven’t seen any other Wikstigators put it quite that way.

Structurally it really does resemble an employment arrangement, albeit an unpaid, uncredited and ostensibly thankless one, where you can quit and nobody would even notice.

Wikimedia Enterprise makes no secret of packaging high-volume structured access to the output of this free labor and selling it to companies that are looking to consume it at scale. In January of this year it publicly named Amazon, Meta, Microsoft, Mistral AI and Perplexity as partners, alongside existing relationships with Google, Ecosia, Pleias, Nomic, Reef Media and ProRata. Some of those arrangements had been quietly formalized over the previous year. Amazon and Meta had apparently been partners for longer but had simply never been publicly identified as such, I guess because they don’t actually have to tell anyone anything if they don’t want to. All this means that there were negotiations and contracts, while the volunteer who wrote the sentence was absent from every stage of the exchange. Most will never know those agreements even exist or where their work ended up, and by the time they decide to remove themselves from the narrative, very little changes — the check has cleared and the article has already been copied into countless downstream systems.

And I’m not saying there’s anything improper about selling access to a public dataset, we all do what we gotta do. While the Foundation has every right to support itself, the issue is that the practical bargain changed *after* the labor had already been performed. A volunteer writes a sentence assuming that it would sit on a page with a talk page and an edit history, but that sentence is now being sold into systems that strip away both only to parrot it back in a confident voice with little visible trace back to its origins. They volunteered for an encyclopedia, but ended up working for an inference engine.

And the subjects have even less say, and no way to opt out.

Meta for example turns the trust level in the RedNipples army all the way up. Its AI research treats Wikipedia as the common knowledge source for benchmarking fact-checking and question-answering systems, and has trained verification models directly on millions of encyclopedia’s editorial choices. So a sentence that may begin as one anonymous volunteer’s contested judgment call, itself a claim meant to be tested, now arrives at the other end as the answer key.

Another irony is that English Wikipedia now prohibits editors from using LLMs to create or rewrite article content (it does permit limited uses such as copyediting and human-reviewed translation), all while actively supplying those same models with its output.

Volunteers may not use AI, but AI may absolutely use them. WP:Don’t get high on your own supply.

All of that, by the way, was decided by a total of 46 volunteers who happened to show up for the open discussion that week. 44 voted in favor and 2 opposed, out of roughly a quarter of a million total active editors.

The loss of all the informational accoutrements becomes more consequential when you try to challenge the result. For all its faults, Wikipedia does give you a semblance of procedure. Even though results aren’t promised, you can certainly try disputing anything you want, and even if you lose (which you likely will), at least you get the satisfaction of leaving behind a record of disagreement for the next person to hopefully find.

You can watch a sentence appear and disappear, return in slightly different form, then get protected by an administrator at three in the morning because somebody had the patience to keep arguing for it.

The messiness is good, it’s the best evidence that the article was written by real people who disagreed. And whatever opinion I might have of the “verifiability, not truth” standard, the important part is that the argument is public and durable by default. It may eventually be archived or removed, but it rarely disappears without leaving a trace.

What the AI model returns is the same disputed sentence, now in immaculate prose, with no indication that five people objected, three got worn down, one was blocked, and the last one standing had the weekend free and an unwavering attachment to the manual of style. A generated sentence has no equivalent procedural address, and it may not trace cleanly to any one source —scratch that — it may not even appear in the same form five minutes later. Ask the same question twice and you may get two different answers.

Admittedly, many AI answers now come with reference links, which does help, but few readers will care to excavate far enough to reconstruct the editorial judgment underneath them. For almost everyone, the footnote will be where the journey ends.

Wikipedia co-founder Jimmy Wales, while making the case for human oversight and transparency, also concedes to the other side of it. “Wikipedia is very transparent, but in some ways it’s so transparent that it’s opaque,” he has said. If you actually want to know why an article says what it says, the answer may be buried under a hundred pages of talk-page debate. “Who has time to read 100 pages?”

He also doesn’t seem to be especially worried about AI removing human attention from the equation. Asked what would happen if Grok took 30% of Wikipedia’s traffic, Wales said it would probably hurt the Foundation’s revenue but not materially change the community: “We’re just a bunch of geeks who like writing an encyclopedia as a hobby.” They would “carry on, even if at a different scale.”

His view is that the traffic disappearing into the big AI is mostly incurious quick-answer visitors, while readers who actually want to go deep will still come.

That may well be true, the encyclopedia might survive without those clicks just fine. What it loses is the range of people who might catch an error before the answer travels everywhere else.

In Part II, the citogenesis ouroboros still required multiple links in the chain. It involved Wikipedia making a claim, some lazy publication repeating it, with the same quote being later presented back to Wikipedia as independent support, and voilà — the snake had successfully gotten hold of its tail. Now AI removes the inconvenience of all these intermediate steps. With scrapers aggressively reproducing whatever is easily accessible and the output swiftly indexed, summarized and regurgitated, a claim that used to require a credulous reporter and an entire news cycle can now pass through thousands of synthetic mouths in an afternoon and come home dressed as corroboration.

It is now both information and authority laundering, and Wikipedia is where everyone gets their papers.

Ruwiki (which I described in Part IV) is what the laundry looks like when somebody removes the pretense: copy the Russian-language encyclopedia, revise the politically sensitive judgments underneath, then place an answer layer on top so the reader meets only the result. The editorial decision is still there, it’s simply been moved somewhere the ordinary person will never see.

Russia had to build an entire fork to do that deliberately, an undertaking that was likely expensive, traceable and painfully obvious. The wider AI ecosystem can perform a version of the same trick everywhere and without asking anyone’s permission.

There is a popular page called Wikipedia:Lamest edit wars, which catalogues the disputes that consumed the greatest editorial effort to the least apparent effect.

Is Hermione Granger a catgirl?

More than forty thousand words on the second i in aluminium.

Eight years on whether yogurt takes an h.

Forty thousand bytes of argument about a high-profile victim of citogenesis — Microsoft’s Clippit, or Clippy.

Six years of edit-warring over whether a snake’s range included “Palestine” or “the Palestinian territories,” with the page noting that the snake’s own views on the political status of Palestine remain unknown. And whatever one thinks of the underlying question, note that Wikipedia itself filed the dispute under “lame edit wars”.

It’s one of the most revealing documents the encyclopedia has produced about itself because it shows that the process has exactly one gear. The same escalation ladder of reverts and noticeboards and requests for comment and solemn invocations of procedure can be applied to a vowel as well as to the opening sentence of a living person's biography.

Sure, there are special policies like WP:BLP but in practice I’ve seen them being applied selectively, if at all.

The cherry on top is the guideline at the beginning instructing contributors not to identify the editors involved in these wars. People have lapses in judgment, it explains, and should not be “stuck with that for the rest of their on-wiki careers for no reason”. Isn’t that sweet of them: an explicit well-reasoned principle against defining a person permanently by a single lapse in judgment. Too bad they only apply it to each other.

And I can’t not bring up Steven Pruitt, the record-holding Wikipedian with more than six million edits under his belt, who avoids biographies of living people whenever he can, partly because the stakes are too high, partly because of general aversion to drama. His preference is for biographies of what he calls “fairly obscure dead people”, because, as he puts it, “they are settled”.

Random thought: if even Wikipedia’s most prolific editor treats a living person’s biography like hazardous material, perhaps more of his colleagues should share that caution.

Meanwhile, the readers have been leaving.

Pew tracked 68,879 Google searches from 900 American adults. When an AI summary appeared, users clicked a conventional result in 8% of visits. Without one, 15%. The source link inside AI summary has been clicked in just 1% of visits.

One percent! Wikipedia is turning into a one-way street with a donation banner, read by hundreds of millions, yet seen by almost no one.

The Wikimedia Foundation reported in October 2025 that human pageviews had fallen roughly 8% compared to the same months a year earlier, after improved bot detection revealed that much of the assumed human traffic was not human at all. It attributed the decline partly to generative AI and social platforms.

LLMs are consuming Wikipedia more than ever while actual people are visiting it less and less, which creates a funding problem as well as epistemic one.

There is no central desk reviewing every Wikipedia article, human readers were always its primary error-correction mechanism. Someone notices their hometown’s population is wrong, or that a BLP’s opening line is a smear, and fixes it, or at minimum complains loudly enough that someone else does. Readers are also the recruitment funnel, since every editor was a reader first. Cut off the readers and you cut off both, at exactly the moment editor and admin numbers are already falling.

The Foundation's own research found monthly active administrators declining since 2018 on more than half of the large Wikipedia editions it studied, while the top 15% generally performed the majority of administrative actions. Unregulated conflict and harassment cause some to reduce their workload or consider leaving altogether. Fewer hands remain, with each one exercising exceedingly more influence over information that can reach vastly more people than will ever visit the article itself.

Same problem with the money: the donation banner only works on humans. A reader who gets the answer from a search summary never sees the ask.

And even though, as mentioned earlier, the Foundation created additional revenue streams by getting Amazon, Meta and co to chip in, that doesn’t restore the lost relationship with its readers community, it merely replaces donors with customers.

And they say I’m only concerned about myself.

There are certainly some solid arguments in favor of pseudonymity. Wikipedia editors have been threatened, sued, doxxed, harassed, arrested, leaned on by governments — like in Belarus and Russia, where contributors and Wikimedia figures have faced state pressure and detention. In India, the Delhi High Court ordered the Foundation to disclose identifying information for volunteers behind the disputed ANI article. Editors in perfectly cushy democracies may avoid contentious subjects simply because they would rather keep their day jobs. The Foundation’s own research found that nearly half of surveyed English Wikipedia administrators had experienced abuse or harassment at least occasionally.

All that is a good reason to protect volunteers, but not good enough to cling onto a system in which public anonymity is also disconnected from institutional responsibility, because Wikipedia often seems to treat the two as though they were the same thing.

Take traditional journalists, who for the most part survive just fine. A newspaper can keep a source confidential from the public while the publisher remains visible and legally accountable for what it prints. The source stays hidden because the institution does not.

And Wikipedia is obviously harder to govern at this scale, but it’s not impossible, and nobody has really tried. I’m not suggesting swapping usernames for home addresses, but what exactly would be wrong with a byline in the institutional sense, where somebody somewhere knows who exercised the editorial power, and answers for the outcome?

On October 3, 2025, the chairman of the Senate Commerce Committee Ted Cruz sent the Wikimedia Foundation a letter alleging ideological bias and demanding evidence. Political diagnosis aside (this post is already long enough), one of his requests was for documents showing what “supervision, oversight, or influence, if any” the Foundation exercises over the editing community. Which is incidentally the same question I’ve been asking throughout this series.

And he got pretty much the same answer everyone gets. The Foundation pointed to volunteer oversight and existing safeguards, reiterated that Wikipedia seeks to inform rather than persuade, and explained that it provides technology and legal support but does not write or control Wikipedia’s content. Its lawyer told Ars Technica that much of Cruz’s letter reflected a misunderstanding of how Wikipedia works. Which checks out — his main mistake was assuming there was someone in charge.

Unfortunately, when there’s no obvious front door, people with enough power will eventually decide that a battering ram is the best way in, while everyone else gets a talk page and instructions to work it out with the volunteers.

Just give people a real grievance officer! Or at minimum, create a point of accountability that is accessible to everyone. But whatever you do, please don’t make it another AI bot.

Can’t resist throwing in some Arendt (again)— her 1967 essay “Truth and Politics” argued that factual truth depends on witnesses, documents, testimony and institutions capable of preserving them. You cannot arrive at it by reasoning it out, and unlike a theorem, a historical event cannot be rediscovered by thinking hard enough. Once the record is gone, it is gone. That’s why societies build institutions meant to stand outside the ordinary struggle for power, like courts, universities, journalism, archives, each charged with keeping the record by some method other than whoever happens to be winning.

Wikipedia has now ended up sitting in that chair, and it was never designed for the weight.

TL;DR: a list of friendly suggestions that I think would make the world a better place, in ascending order of how much the editors are likely to hate them.

1. Make the provenance travel with the sentence. Package the dispute and revision metadata into the feeds. The companies buying those feeds should keep it attached when they generate a claim about somebody, and the disclosure should be reachable by the person reading the answer. A sentence shouldn’t arrive as a settled fact about a person when it is really the surviving outcome of an argument between strangers.

2. Add public article scoring that readers can actually see. Wikipedia already does some grading — stub, start, C, B, good article, featured — but that scale measures completeness and craft, lives on the talk page where nobody goes, and is assigned from inside the same volunteer pool. I want the other axes rated: accuracy, fairness, neutrality, and how contested the article currently is. Displayed at the top of the page, not buried one tab away. An article that eleven people have fought over for six years should not present itself to a reader with the same untroubled face as one nobody has touched since 2009.

3. Create an independent appeals office for biographies of living people and similar consequential articles, one that doesn’t involve the volunteer who made the edit that’s being appealed. Deadlines. Authority over sourcing, weight, attribution, labels, omissions, images, and the factual disputes nobody resolved. Written reasons, and an identifiable person on every ruling. Power to order removal, qualification, attribution, or correction when the ordinary process has failed. Let the subject attach a clearly labeled comment that’s visible in the main body in case of a dispute.

4. Create a narrow statutory carve-out from Section 230, which currently protects the Foundation from being treated as the publisher of material supplied by volunteer editors, largely foreclosing defamation claims against it in U.S. courts. I would draw the exception tightly, it should apply only to specific credible claims that a factual statement about a living person is defamatory. The complainant would have to identify the exact language, explain why it is materially false, and use a formal notice process. Wikimedia would then have a fixed period to investigate, correct or reject the complaint with reasons. Review it in good faith and immunity remains. Ignore it, or knowingly leave a clearly actionable falsehood in place, and immunity for that claim is lost. That wouldn’t mean Wikimedia loses the case, the claimant would still have to prove every element of defamation, and public figures would still face the actual-malice standard. Many claims would fail on the merits, as they should. Right now, responsibility disappears at every layer: the editor may be unidentifiable, the Foundation is shielded from publisher liability, and the model downstream repeats the claim without answering for it. A narrow exception would create one circumstance in which “nobody is responsible” is no longer the final answer. And I did consider the objection that lawyers would weaponize the process. Some would certainly try, but it’s not like they aren’t already doing it. The current arrangement does not prevent legal intimidation, it only ensures that it arrives in form of political investigations, foreign judgments, emergency motions, and threats to block the whole site. A front door is less dangerous than a permanent siege. Build it with a narrow standard, evidence requirements, penalties for abusive notices and protection for editors acting in good faith. Then staff it with somebody who can throw out a meritless demand without six months of ceremony.

5. Allow a first-person perspective in biographies of living people. Not the right to rewrite the article about yourself, but a disclosed, clearly labeled statement from the subject, attached to a disputed passage, for readers to weigh themselves. This would be most useful on the uncontroversial stuff, for example a biographical fact that’s been wrong for years and keeps getting copied outward. There is no meaningful difference between a subject giving a quote to a reporter and saying the same thing in the open under their own name. The current arrangement is that a living person must persuade a reporter to quote or write about them, wait for publication, then bring that article back to Wikipedia before their own account becomes admissible. The detour buys a reporter's judgment about what was worth publishing, and it’s valuable in a lot of cases. But once the subject is notable enough to warrant an existing Wikipedia article, the question is why their own account still counts for nothing. The press is not a neutral filter either, and can be worked by anyone who knows how. The prospect of original content would also give people a reason to visit Wikipedia directly.

6. Add verification to gain elevated permissions required to edit a certain tier of article— like biographies of living people, current political events, and similar consequential pages. By verification I do not mean handing over your name to the public. The editor should be privately known to the Foundation, or to an independent body set up to hold exactly that information. Verified once, pseudonymous to readers, answerable to somebody who can take the privilege away. Editors facing documented danger get legal support, security help, and whatever else the project can afford. Absolute anonymity is rarely necessary, and accountability tends to arrive far too late. Knowing that anonymity is conditional rather than absolute does more to shape behavior than most written rules. Wikipedia’s existing guidelines already recognize that a biography of a living person requires different treatment from an article about let’s say a root vegetable. Maybe the editing permissions should reflect that too, instead of blindly hoping everyone will follow the rules. Anonymity can be a shield, but it should not become a throne. And before someone says this would end the volunteer model: Wikipedia is already losing editors as is, and the remaining ones end up wielding disproportionate power as a result. Some might feel compelled to leave. Others will show up precisely because there is finally a structure worth joining…

I started this series with obviously selfish motives, but I never considered that a disqualification. If anything, Wikipedia’s current structure gives the editor considerably more procedural power than the subject, and conflict of interest does not disappear just because it isn’t in-your-face obvious from one side. The alternative rule —that only people unaffected by a system may propose changing it—is an extremely convenient one for whoever is doing the affecting.

Žižek’s version of this, in The Parallax View, is that universality never sits in some neutral container above the fight, it surfaces in the element the structure has no place for. Sometimes being implicated gives you the better view.

As for the original request that started this entire thing, I no longer have to tell you it was denied. At least not as of the publication. I have no idea how long it’ll hold, everything might be reverted shortly after someone reads this (but pls don’t!).

All I can say is that for now I’m continuing to monitor the situation. It’s also been a pleasant surprise to see how many people reached out eager to bond over shared suffering. And in the end, isn’t it all about the friends we make along the way?

Share

Thanks for reading. If you made it this far —wow. I do wonder what went wrong in your life. And if this series was useful to you, please subscribe and share and don’t forget to nominate me for Wikipresident.

Subscribe

This is a series about Wikipedia’s anonymous editorial power — and what happens when AI turns contested human judgments into facts.

Don’t miss Part I, Part II, and Part III.

Share

You have biographers.

You did not choose them, and you cannot fire them. They have never met you, and they may know nothing about your work, your life, or the subject they are editing, but this is apparently irrelevant. They work for free, under pseudonyms of their choosing, inside a website that now supplies the first draft of reality to Google, journalists, schools, and artificial intelligence. You are not entitled to know who they are. They, meanwhile, are allowed to decide who you are.

Very normal system.

The names are almost the only thing missing. A remarkably detailed portrait of the profession exists because Wikipedia, with the self-documenting compulsion of an institution that files paperwork on its own pathologies, surveys its editors and publishes the results. The latest put women at roughly 13% of seasoned editors and just 7% of administrators. Older studies placed the male share of the editing community as high as 90%. The numbers wobble between studies and methodologies, but they wobble between “overwhelmingly” and “almost entirely.” Whatever else can be said about the world’s biographers, they are certainly not a representative sample of the world.

The output looks exactly like the input.

As of August 1, 2026, English Wikipedia contained 2,142,296 biographies. Of these, 436,909 — 20.39% — were about women. One in five. Not one in five contemporary figures, where you might blame the imbalance entirely on history; one in five, total, after two decades of the most ambitious biographical project ever attempted, conducted by a community with access to an unprecedented share of the world's digitized record.

There are volunteers working on it. In 2014, researchers quantified what women editors had long suspected: just 15.53% of English Wikipedia's biographies were about women. In July 2015, Rosie Stephenson-Goodknight and Roger Bamkin established a project called Women in Red, named for the red links that mark articles which do not exist. On Wikipedia, blue means someone thought you mattered. Red means the encyclopedia knows your name only as an absence, a placeholder waiting for a biographer who has not shown up.

Women in Red exists to turn those red links blue, one missing woman at a time. Its members have helped create hundreds of thousands of biographies the ordinary operation of the community somehow never got around to. After more than a decade of sustained volunteer labor, the proportion rose from 15.53 to 20.39%.

This is presented as progress, largely because the previous number was even worse, which is also how one describes the successful management of a chronic disease.

Wikipedia did not invent the exclusion of women from history, journalism, science, or public life. Its defenders correctly point out that an encyclopedia can only summarize the sources that exist.

True, not a defense.

Once an institution converts inherited undercoverage into a present-day test of notability, it no longer merely records historical exclusion. It gives that exclusion administrative force. Women received less coverage because they were barred from the institutions newspapers and universities considered important. Their lack of coverage is then cited as evidence that they do not qualify for Wikipedia. Their absence from Wikipedia reduces their discoverability, making future coverage less likely. The archive supplies the prejudice, the policy then notarizes it. The bias enters through history and exits as procedure.

The world did the excluding first. Wikipedia arrived later to alphabetize it.

If my biography is the wrong test case, pick another. Donna Strickland will do. The argument survives without me, I merely happen to have done the fieldwork.

In March 2018, someone submitted a draft article about Strickland, an optical physicist whose work on chirped pulse amplification had already been foundational to the field for three decades. A reviewer declined it that May. The references, the verdict explained, did not demonstrate enough significant independent coverage to establish that the subject qualified for an article. Translation: the work may have transformed modern physics, but too few newspapers had said so.

Five months later, Donna Strickland won the Nobel Prize in Physics — the first woman to do so in fifty-five years. At the moment the prize was announced, she had no Wikipedia article. Gérard Mourou, who shared the prize with her, had had one since 2005. Strickland's page was assembled in a panicked ninety minutes after the Stockholm press conference, while the encyclopedia of record discovered, in real time, a woman whose draft it had formally found unsupported by sufficient independent coverage just months earlier.

The defense arrived immediately: the draft was thin, the sources were insufficient, the reviewer followed the rules, the process worked as designed.

Precisely. The process did work as designed. That is the problem.

Strickland was not overlooked because she was unimportant. She was judged insufficiently documentable because the institutions responsible for manufacturing public notability had paid insufficient attention to her, and Wikipedia mistook that neglect for an objective measurement. The bureaucracy did not malfunction. It processed the bias exactly as instructed.

A future Nobel laureate, rejected in advance by the bureaucracy built to overlook her.

Notability, it turns out, is not a property of the person. It is a property of the audience.

And the machinery was not always this skeptical.

Until 2022, Wikipedia’s football notability guideline treated a single appearance in a competitive match between fully professional teams as evidence that significant coverage was likely to exist. Start the match or come on as a substitute: either would do. One professional appearance and the presumption switched on.

Strickland’s three foundational decades in optical physics did not satisfy a reviewer. Forty-five minutes of professional football could satisfy a written rule.

This was not merely a thousand editors independently preferring football. The preference had hardened into written guidance. The encyclopedia had taken one community’s enthusiasm, converted it into a bright-line test, and taught the bureaucracy to recognize a team sheet more readily than a scientific legacy. The participation rule was finally removed after a 2022 discussion concluded that playing one professional match did not, in fact, guarantee that enough independent sources existed to write a proper article.

The institution eventually noticed the absurdity, but only after filling the encyclopedia with it.

The skew does not end with sex. The editing community is disproportionately drawn from highly educated, urban populations in the wealthier and better-connected parts of the world. In Wikimedia's 2024 survey, 81% of active editors had completed postsecondary education, 42% held postgraduate degrees, and 61% lived in a major city or metropolitan area.

That would matter less if English Wikipedia were merely one publication among many. It is not. It is the flagship: the largest, the most consulted, the version most likely to surface first in search results and, lately, to enter the machines as reality's default draft. The whole world is described in it. A very small slice of the world does most of the describing.

And attention goes where the editors already are.

Nearly every episode of The Simpsons has its own article: plot summary, production history, cultural references, critical reception, you name it - all meticulously sourced and vigilantly maintained.

As of mid-2026, English Wikipedia contained roughly 790 to 800 individual episode articles, out of 808 episodes aired — nearly one dedicated article per episode. Roughly eight hundred articles for one cartoon.

Nor does the appetite stop with television. World of Warcraft has a principal article, a separate article for each of its eleven expansion packs, and standalone entries for everything from its fictional characters to a contagious virtual disease and a player who became famous for ruining one raid.

And the pattern is not confined to what airs or ships. Researchers at Oxford found more geotagged Wikipedia articles about Antarctica than about any country in Africa or South America, and more about the fictional places of Middle-earth and Discworld than about many actual countries across Africa, Asia, and the Americas. The Second Congo War and its aftermath, associated with an estimated 5.4 million excess deaths, the deadliest toll attributed to any conflict since 1945, gets one principal article. The Simpsons gets nearly one for every episode.

Wikipedia does not lack space. It has preferences.

Subscribe now

This is not an argument that Homer Simpson deserves less documentation. I would never want to interfere with a man’s sacred duty to preserve the production history of “A Hunka Hunka Burns in Love.” It is an observation that Wikipedia’s attention follows enthusiasm, and enthusiasm is not a synonym for importance.

Entire regions of the world — their cities, literatures, political histories, artists, and living scientists — remain thinner on Wikipedia than a fictional suburb’s Halloween specials. Nobody voted to prioritize Springfield over the Democratic Republic of the Congo. No committee convened to decide that the cultural references in season fourteen required more urgent preservation than the intellectual history of Central Africa. No decision was necessary.

That is almost worse. The imbalance emerged automatically, no malice, no design, no sinister meeting in a dark room. Just thousands of individual preferences accumulating until preference became structure and structure began presenting itself as reality.

This is what "anyone can edit" looks like in practice.

It means anyone with broadband. Anyone with leisure time. Anyone with fluent written English. Anyone with access to sources the existing community recognizes. Anyone with enough social confidence to survive an afternoon of strangers explaining why the subject they care about does not count.

When that pool is heavily skewed by gender, geography, education, and language, the encyclopedia's map of the world skews in exactly the same directions. Wikipedia's defense is that nobody planned it. Correct - it’s not a conspiracy, it’s an outcome.

History may be written by the victors. Wikipedia is written by whoever had Saturday free.

The coverage is the symptom, the composition is the condition. The encyclopedia does not reflect what exists, it reflects who showed up.

Which brings us back to the question I left hanging in Part III: who are these people, and why can't we know?

I promised we’d get to the more exciting countries, and here they are — because honesty requires admitting that this is where the case for anonymity stops being theoretical and starts collecting prison sentences.

In March 2022, pro-Kremlin Telegram channels published the identity of Belarusian Wikipedia editor Mark Bernstein and accused him of editing articles about Russia's invasion of Ukraine. Belarusian authorities arrested him, and he was initially sentenced to fifteen days' detention and later to three years of restricted freedom.

In Saudi Arabia, two longtime Wikipedia administrators were imprisoned after contributing material about political detainees. One was sentenced to eight years. The other was initially sentenced to five years, had that sentence increased to thirty-two years on appeal, and later saw it reduced — eventually to fourteen years in September 2025.

Russia, tiring of pressuring individual editors, eventually chose a more efficient instrument - it forked the entire project. Ruwiki, a state-supported copy of Russian Wikipedia, began with approximately 1.9 million articles from the original and then set about removing or rewriting inconvenient material. The invasion became a "military operation." Occupied territories were linguistically laundered, dissidents were trimmed, state criticism was reduced. The whole thing was then promoted as a domestically compliant alternative.

A digital gulag with the original's fonts.

Leave a comment

So here is my point. For an editor in Minsk or Riyadh, anonymity is not a perk. It can be the difference between editing an encyclopedia and entering the criminal justice system. Any identity regime that simply published the legal names of vulnerable editors to the world would be not merely careless but obscene. If Wikipedia's anonymity policy were a calibrated shield, protection scaled to threat, I would have nothing to say against it.

But it is not a calibrated shield. It is one-size-fits-all body armor distributed without regard to who is actually being shot at. The entitlement to public pseudonymity is effectively universal and largely indifferent to the level of danger. The editor facing an actual secret police and the editor in a liberal democracy facing ordinary reputational discomfort receive the same public-facing protection. The Minsk case is invoked to defend the Milwaukee hobbyist, and the moral prestige of political persecution is quietly distributed to everyone with a handle.

An editor in a stable democracy can point to a dissident in prison and say: this is why you may not know who I am. A very convenient arrangement for that editor, and less so for the person being described. There is something theatrical about a system in which editors working safely in liberal democracies wrap themselves in the security requirements of dissidents whenever an article subject asks who exercised judgment over her life.

Worse, the shield is weakest where it is most needed and strongest where accountability is most ordinary. Pseudonymity can raise the cost of identifying an editor. It can deter casual harassment, delay exposure, and protect people from opportunistic retaliation. It cannot guarantee protection against a determined government. States have surveillance systems, infiltrators, subpoenas, device seizures, telecommunications records, border controls, and employees whose professional responsibility is discovering who you are. Bernstein was behind a username, so were the Saudi administrators. A universal rule is a poor substitute for protection calibrated to actual risk.

Against a determined state, the mask rarely holds. Against an ordinary person asking who selected the insult in the first sentence of her biography, it is impenetrable.

Tissue paper facing power. Armor facing accountability.

Notice what this geography of danger does to the encyclopedia itself. The places where editing carries the greatest personal risk are, not coincidentally, among the places with the thinnest and most distorted coverage. The regions that most need documenting are the ones where documenting can be dangerous, so the work is done, when it is done at all, by outsiders, from elsewhere, often in English, applying sources and concepts produced somewhere else again.

Which is the strongest possible evidence that blanket anonymity is not working as protection. It is granted universally, and it delivers least where it was needed most. The encyclopedia knows most about the people least at risk of writing it.

Share

So, returning to the safe countries: the journalist comparison from Part III is worth finishing, because the incredulity was only half of it.

Apparently, choosing to wield power over living people for free entitles you to more anonymity than the professionals paid to perform comparable work.

This is usually presented as self-evident: they are volunteers, they might be harassed, they might quit. Both things may be true, neither answers the question. The absence of compensation somehow functions as both a credential and an alibi.

To be clear, newsrooms are not innocent of selective attention, they carry their own severe biases, and they, too, can treat entire regions as intermittent crises rather than continuous realities. The difference is not that newspapers are pure while Wikipedia is compromised, the difference is structural. A journalist's accountability was never contained in the paycheck, it came from the masthead, the assigning editor, the publisher, the correction process, institutional reputation, libel lawyer if it comes to that. All the structures capable of attaching a real consequence to a real decision made by a real person.

Wikipedia, of course, has administrators, noticeboards, arbitration committees, conduct policies, and enough procedure to qualify as a minor principality. It does have process, but what it does not have is the ordinary chain of named editorial responsibility- there’s no assigning editor, no accountable masthead, no named publisher standing behind the individual judgment, no institution to which the subject can submit a correction and receive a decision from someone whose own name appears on it, and no consequence that can reliably travel beyond the pseudonym.

The Wikimedia Foundation provides the infrastructure while disclaiming control over editorial content. The volunteers control the content while disclaiming the obligations of professional publishers. Authority is dispersed while responsibility evaporates.

Hannah Arendt, writing about bureaucracy in On Violence, called it “the rule by Nobody” — a structure in which “there is no one left who could even be asked to answer for what is being done.”

Wikipedia is not a state, and its editors are not Arendt’s bureaucrats. But the structural resemblance is difficult to miss - power is distributed so widely that accountability disappears between the participants. An article remains, naturally, everyone’s responsibility in theory and nobody’s in particular. A perfect circle- Wikipedia does love one.

And unlike any newsroom, Wikipedia has the scale and default authority of the version most likely to surface first and to train the machines, while remaining almost impossible to hold to account.

Reach of a publisher, posture of a hobby, responsibility of neither.

Remove both the salary and the structure, and what remains is power with nothing attached, exercised over living people by strangers who treat the absence of a paycheck as a form of immunity.

Every role involving consequential judgment about other people carries some degree of risk. Journalists face harassment, judges receive threats, prosecutors are great at making enemies. Professors, politicians, doctors, and corporate executives all attract retaliation from people who dislike their decisions. We generally do not respond by erasing their identities. We respond by expecting institutions to protect them while preserving accountability. So why should Wikipedia be different?

And if the answer is that it cannot be different, that the role genuinely cannot be performed with any meaningful accountability because the personal risk is deemed unacceptable, then perhaps the question is no longer how much anonymity the role deserves. Perhaps the question is whether private volunteers should be exercising that level of authority over living people in the first place.

That is not an attack on every editor's motives, and it does not deny that harassment occurs. It does not require publishing every editor's legal name, employer, address, or personal information to the world. Accountability and public exposure are not synonyms. Accountability means that somewhere beyond the username, a real institution knows who is exercising elevated power, establishes what that power permits, and accepts responsibility when it is abused. Someone who knows the name, sets the rules, and answers when the rules fail.

At present, the subject is visible, searchable, indexed, summarized, and fed into machines, while the editor remains a cartoon avatar and a string of characters.

The question was never whether volunteers deserve privacy, it is: what level of accountability should accompany this level of power?

Part V will attempt an answer.

In Part V, the finale: what the machines learned from the anonymous court and why they can never unlearn it, where Wikipedia's readers went, and a few modest proposals the editors will hate — including one that involves lawyers.

This is a series about Wikipedia's anonymous editorial power — and what happens when AI turns contested human judgments into facts.

Don’t miss Part I and Part II.

Share

Subscribe

So who are the unpaid boy scouts guarding our civilizational artifact?

With more than 7.2 million articles, English Wikipedia has roughly 30,000 to 45,000 “active” editors in a typical month, with “active” meaning at least five edits within the preceding thirty days. Anything from creating an entire page to adding a comma counts as an “edit.”

Then there are the admins—the volunteers entrusted with the ability to block accounts, protect pages, delete material, perform clerical miracles by which disagreement quietly ceases to exist, along with various other acts of procedural euthanasia. As of this writing, English Wikipedia’s active administrators number 435, by the project's own threshold of thirty edits in two months.

Officially, being an admin doesn’t give one’s opinion more weight in a content dispute, Wikipedia's stated principle being that adminship is "no big deal," a maintenance role, not a rank. Formally, a maître d’ merely holds the reservation book. Practically speaking, see who gets seated.

Administrators have, after all, survived a community approval process, accumulated institutional standing, mastered the local dialect, and obtained powers unavailable to ordinary users. This does not make every administrator a tyrant, but it does make the claim that the role carries no additional influence feel like the sort of constitutional fiction institutions recite immediately before doing exactly what they were supposedly powerless to do.

Four hundred and thirty-five people, give or take, quietly holding the administrative keys to what has become the default memory of the English-speaking internet.

I am not claiming that every single one of them is corrupt, malicious, stupid, megalomaniac, or clinically overinvested in capitalization disputes. I have certainly encountered editors who were fairly helpful and who did a great job convincing me they were acting in good faith. Many volunteers have sacrificed what I can only imagine are grotesque amounts of free time trying to create something genuinely useful. Wikipedia would not exist without them. (And what a shame that would be.)

But gratitude is not diplomatic immunity, and it is certainly not a wild suggestion that a small, self-selecting group exercising extraordinary practical influence over an information system relied upon by so many should be held to some standard of accountability.

Besides, the people with the greatest influence over an article are often neither administrators nor leaderboard superstars. They’re simply whichever few pseudonymous enthusiasts have placed the page on a watchlist and developed an emotional commitment to police and defend it with the intensity of a medieval border dispute. An obscure article can become the private allotment of whichever editor has the strongest territorial instinct and the least competition from ordinary life.

That is how power works on Wikipedia: locally, procedurally, and through total exhaustion.

Participation is radically unequal. Of course, not every edit represents serious editorial judgment. Many are formatting changes, vandalism reversions, automated corrections, category adjustments, and other valuable forms of digital janitorial work. But the overall structure is obvious: a minority of extremely active accounts shoulders most of the working load, learns the procedural maze, accumulates status, recognizes the regulars, and acquires far more practical authority than the phrase “anyone can edit” would ever lead a normal reader to suspect.

Some of the regulars seem to experience Wikipedia less as an encyclopedia than as World of Warcraft for people with strong opinions about citation formatting. There are levels, badges, permissions, enemies, alliances, obscure rules, ceremonial language, and endless side quests.

Share

Tens of millions of registered accounts create the appearance of mass participation. A few thousand highly active users do much of the meaningful governing. Hundreds of administrators possess the serious enforcement tools, and fewer than that remain consistently active. Any one of Wikipedia’s seven million articles can become the rent-controlled apartment of whoever arrived first and never emotionally moved out.

Concentration itself is not the concern. In the end, every serious institution relies on a smaller group of experienced people. The concern is concentration without meaningful identity, qualification, disclosure, or accountability. The way things stand today, the largest database of human knowledge is being governed by a relatively small corps of unelected, largely pseudonymous accounts whose employers, affiliations, financial arrangements, ideological commitments, personal grievances, motives, and subject-matter expertise are unknown to the public, unless they choose to disclose them, in which case the public is expected to take their word for it.

Speaking only of liberal democracies for now (we’ll get to slightly more exciting countries later) - I genuinely struggle to understand the insistence on absolute anonymity. Journalists attach their names to investigations of organized crime, political corruption, billionaires, governments and conspiracies every day, yet I’m supposed to believe that correcting the punctuation on the List of Byzantine Emperors requires witness protection.

Wikipedia’s main defense is that its decisions are made not through credentials, hierarchy, or institutional authority, but through sources, policy, discussion, and consensus. And when the process works as intended, it can look remarkably conscientious.

In 2022, when partisans descended on the article for “recession” and attempted to conscript its definition into a political messaging war, editors protected the page, pushed the fight onto the talk page, and eventually produced language more restrained than most of what was being pushed by professional commentators.

When Elon Musk's gesture at a 2025 rally was compared to a Nazi salute, editors spent roughly 7,000 words deciding how to describe it without claiming to read his mind: the physical motion was stated, the comparison was attributed, and Musk's denial was included. Thousands of words entered the system, consensus was reached, three sentences survived.

Wikipedia has also generated more than 40,000 words debating whether the "I" in Star Trek Into Darkness should be capitalized. Consensus was reached there, too. The letter remained uppercase.

All three are examples of Wikipedia working as intended — which may be the best illustration of both its virtue and its pathology. The same machinery, the same priestly vocabulary, and the same procedural gravity are applied to economic history, an alleged Nazi salute, a living person's reputation, and one capital letter in the title of an otherwise forgettable Star Trek movie.

Still, these are unusually visible disputes, or at least unusually attractive ones to the type of person prepared to spend several working weeks defending a vowel. But most articles do not receive this level of communal vigilance. Most disputes are not settled by a representative council of sober-minded editors carefully weighing the evidence. After experiencing the process firsthand, I can confirm that at WikiHQ the word “consensus” is being stretched to its outer limits, because it means very little when nobody involved is obligated to remain until the discussion is finished.

Wikipedia’s editors are volunteers. Nobody hired them. Nobody assigns the case. Nobody can require them to participate, answer an objection, confront contrary evidence, or produce a final reasoned decision. There are no deadlines, no deliverables, and no identifiable editor responsible for ensuring that the process reaches any coherent conclusion at all. Authority is diffuse, responsibility is difficult to locate, and no single person appears answerable for the result. The whole arrangement resembles a Turkish bazaar trying to operate under the rules of decentralized bureaucracy.

A common scenario looks something like this: a stranger wanders into the argument, someone responds, one person cites three policies, another cites five. A newcomer is directed to an archived discussion, then to the archive of the dispute over whether the first archive was interpreted correctly. Acronyms begin multiplying, templates appear, and by the fourth reply, someone cites an essay written in 2009 by another anonymous volunteer explaining why an even older essay should not be interpreted literally except on alternate Tuesdays. Everyone speaks as though The Hague is waiting breathlessly for the ruling.

Then reality begins culling the field. The person with a job leaves, the person with children leaves, the person with a dinner reservation leaves. Someone remembers they own a dog. The person who merely wanted an obvious factual error corrected discovers that the process now requires familiarity with twelve policies, six noticeboards, two acronyms that appear to contradict each other, and the temperament of a medieval canon lawyer. Eventually, the only person left is the one with the greatest procedural endurance and the least compelling reason to be anywhere else. And the silence left behind is recorded as consensus.

But a volunteer editor becoming bored and wandering away is not consensus. Neither is a living subject giving up after being buried beneath procedural debris.

Attrition is not agreement. Silence is not consent merely because everyone else eventually went to bed.

Wikipedia disputes are formally supposed to be decided by superior sourcing and closer adherence to policy. In practice, they are often won by whoever has most fully absorbed the local bureaucracy and can spend the greatest number of hours restating the same position in its preferred ceremonial language.

Hardcore Wikipedians describe this as quality control: outsiders arrive with complaints, Wikipedia patiently explains why the real problem is their failure to understand how Wikipedia works. And occasionally, the outsider remains long enough to learn the rules and participate successfully. But more often, the process simply outlasts them. They return to their jobs, children, dinners and dogs, while the remaining editors interpret their departure as the natural triumph of consensus.

WikiStans like to imagine the project as an immune system. The metaphor is apt, though perhaps not in the way its admirers intend. It raises the obvious question: what, exactly, does the system recognize as the pathogen? Healthy immune systems eliminate pathogens. Autoimmune diseases attack the body they were meant to protect.

A mature editing community can become so optimized for detecting “bad” edits that it starts rejecting legitimate contributions, especially from newcomers or people with conflicts of interest.

Factual errors can remain in an article for years. An unsupported insinuation can become embedded through repetition. A dubious characterization left undisturbed for long enough ceases to be controversial and graduates into orthodoxy, at which point the citogenesis loop described in Part II is ready to take over distribution.

But let the living subject object to it, and the immune response activates immediately. Their motives are questioned, their conflict of interest is announced, their language is scrutinized, their sources are treated with suspicion. Their request is reformatted, redirected, deferred, archived, and surrounded by warning labels.

The error may have slipped through unnoticed. The correction arrives carrying twelve forms of identification.

The system is therefore not always optimized to eliminate falsehood. It is optimized to resist outside interference. And since the person described in the article is treated as outside interference by default, the article eventually stops protecting the subject from error and begins protecting the error from the subject.

An idea: allowing the subjects of biographies to append a clearly labeled account in their own words would cost Wikipedia little, except perhaps the comforting fable that anonymous editors alone are licensed to construct the narrative.

Share

Larry Sanger, Wikipedia’s freshly banned co-founder, helped design the institution’s original editorial philosophy, including its neutrality policy. He later became one of its most persistent critics, arguing that the project had hardened into ideological conformity, concentrated authority in pseudonymous power users, and substituted a very elastic idea of consensus for anything resembling accountable government. 

Sanger’s own record is hardly spotless. His politics are contentious and he is not always the best custodian of his own arguments. Even so, some of his criticisms are difficult to dismiss.

In 2025, Sanger proposed a package of reforms: publicly identifying Wikipedia’s leadership, ending indefinite blocks, restoring what he regards as the project’s original neutrality standard, permitting competing versions of disputed articles, and replacing the warm fog of consensus with a more formal governing structure. The editing community considered his charge that Wikipedia had become insular and responded with the serene unanimity of an institution determined to prove him right.

And—surprise—in June 2026, he was blocked indefinitely. The editors concluded that Sanger had violated Wikipedia's prohibition on off-site canvassing by encouraging outsiders to participate in an internal discussion concerning a proposed WikiProject. The ban was imposed before the mandatory seventy-two-hour discussion period had even run—even Jimmy Wales, the cofounder still in good standing, called it "ludicrous" and briefly lifted it himself—then reimposed once the clock expired: the rulebook violated and cured by the same hands. The debate also invoked the capacious standard known as "not here to build an encyclopedia," a rule malleable enough to acquire whatever shape the people enforcing it require. Sanger disputes both the characterization of his conduct and the fairness of the proceeding.

One need not endorse his politics, accept every accusation he has made, or purchase the complete Larry Sanger reform bundle to find the episode informative.

Although he officially stopped being a part of Wikipedia decades earlier, I found the story eerily familiar: a co-founder of one of the most influential reference works in human history expelled by a self-governing collection of volunteers, many operating behind pseudonyms, through a process in which the people interpreting the rules, identifying the offense, evaluating the evidence, announcing the consensus, and imposing the punishment substantially overlapped. Judge, jury, prosecutor, bailiff, and court stenographer — all signed in under names like LordOfChairs69, CommaSentinel_1998, and of course RedNipples. 

Wikipedia’s defenders will say that this proves nobody is above policy. Perhaps. It also proves that nobody is beyond the reach of those empowered to interpret it.

When legislative, prosecutorial, judicial, and enforcement functions collapse into the same pseudonymous community, asking who these people are is not harassment, and wanting to know what qualifies them is not an attack. Questioning whether they have undisclosed affiliations, grudges, employers, ideological commitments, or conflicts of interest is not an assault on the encyclopedia. It is the minimum curiosity normally applied to anyone exercising institutional power.

Wikipedia expects the public to admire the purity of a system in which authority is everywhere, responsibility is nowhere, and the people administering both may remain unidentified. 

It demands radical transparency from its subjects while extending radical opacity to the people judging them.

Transparency for the defendant, anonymity for the court. 

And now the court’s freelancing for your favorite chatbot.

In Part IV: Meet Your Biographers—an encyclopedia edited by a community that is roughly 90% male, why every episode of The Simpsons outranks entire regions of the world, and how a future Nobel laureate was rejected in advance by the bureaucracy built to overlook her.

This is a series about Wikipedia’s anonymous editorial power — and what happens when AI turns contested human judgments into facts.

Read Part I Here

Wikipedia’s problem begins with the premise that significance can and should be measured by the amount of media coverage a subject receives.

Its editors establish notability through “independent” reporting, which in turn determines which facts deserve prominence, which descriptions become canonical, and which events are allowed to define an entire life. The entire system rests on a theory of contemporary media so naive it borders on the delusional- that journalists cover people and events in proportion to their objective importance, describe them in language finely attuned to reality, and divide their attention fairly between the admirable and the disgraceful. None of this is true.

The modern media economy is less a repository of truth than a carnival barker with an analytics dashboard.

In a rat race over shrinking attention spans, outrage moves and scandal converts. Malice has excellent click-through rates. A decent act might earn three sentences beneath a banner mattress ad, but an accusation gets a headline in a push notification, fifteen rewrites, and a permanent home at the top of Google results. Anyone with even a tentative grasp of human nature understands the imbalance. Bad news beats good news every time. Disgrace is sticky, vindication is merely housekeeping.

And Wikipedia’s anonymous, self-anointed custodians are not nearly as innocent as their sourcing catechism would have it. They solemnly blacklist the Daily Mail as unreliable, then reliably reproduce its worldview: the same appetite for degradation, the same fixation on scandal, the same smug conviction that the ugliest available framing must therefore be the most serious.

Build an encyclopedia on top of that sewage and you do not get a neutral record of human significance, regardless of how acrobatic one’s imagination might be. You get a scandal-ranking engine with footnotes, administered by pseudonyms and dressed in a Greek portico so everyone mistakes the stench for scholarship.

Subscribe

This is not to say Wikipedia is useless, quite the opposite. It is one of the most remarkable collaborative projects ever built, which is precisely why its editorial assumptions deserve scrutiny. The more indispensable an institution becomes, the less we can afford to mistake its procedures for laws of nature.

The distortion does not even require anyone to publish a literal falsehood- every individual sentence may survive fact-checking while the overall portrait remains grotesquely untrue. A biography can contain only sourced facts and still resemble a police blotter assembled by one’s worst high school enemy (speaking from experience). Accuracy at the sentence level does not produce truth at the scale of a human life.

Wikipedia’s policies are exceptionally good at asking whether a claim appeared in an approved publication. They are considerably less interested in whether thirty years of someone’s existence have been compressed into the six weeks most profitable to the attention economy. Wikipedia can verify the brick while ignoring the shape of the building. It can even count the bricks- so many sentences for the accusation, so many words for the reply- and mistake arithmetic for fairness.

And if the worst month of your life generates more journalism than the preceding decade (trust me, it’s not that hard), Wikipedia does not see an imbalance to correct. It sees two hundred reliable sources.

And it gets better.

Media outlets habitually consult Wikipedia, either directly or laundered through search engines and LLMs. You can spot it instantly- their prose begins to sound like our free encyclopedia, sentence by sentence. Then, as the expired maraschino cherry on top, some brilliant logician in an editing mood cites those very outlets as “independent support” for wording that originated on Wikipedia in the first place. The loop closes, the consensus cites itself, and good luck to anyone trying to change the station. What began as the preference of whichever RedNipples arrived first, stayed longest, and possessed the greatest tolerance for procedural trench warfare returns wearing the costume of external validation. Darwinism, if Darwin had selected for free time.

From there the characterization migrates into Google results, TV scripts, podcast introductions, social media biographies, AI summaries, and yet more articles. Each repetition is welcomed as another ostensibly independent confirmation, until every source is a mirror and every mirror is admissible— unless, naturally, its name is the Daily Mail, or The Points Guy.

Wikipedia behaves like a mutant ouroboros with a citation toolbar. The more often a claim is repeated, the truer it appears. The truer it appears, the more weight it receives. The more weight it receives, the more prominently Wikipedia displays it. And the more prominently Wikipedia displays it, the more likely everyone else is to repeat it.

The snake eats its tail and bills you for the tasting menu. Gratuity not included.

The phenomenon even has a proper name — citogenesis — coined, fittingly, not by an epistemologist but by a webcomic, and pointed out to me by D.F. Lovett, who loves writing about Wikipedia even more than I do, if that’s even possible.

Once a characterization enters the machine, Wikipedia becomes its distribution network. It is copied swiftly, widely, and with the confidence of a man selling his How to Lose Ten Million in Ten Days course on LinkedIn. Corrections, meanwhile, move at the speed of a municipal permit application. The original wording can travel around the world before an editor has finished explaining why changing it would be too risky or, God forbid, “controversial.”

When Hannah Arendt wrote that “the difference between truth and falsehood may cease to be objective and become a mere matter of power and cleverness, of pressure and infinite repetition,” she might as well have been writing about Wikipedia, which now sits deep inside the training data of every major language model. AI has obsequiously volunteered to industrialize the infinite repetition.

Wikipedia no longer needs to persuade humans, it only needs to enter the machines. Once the machines repeat it, humans mistake the echo for corroboration.

Corroboration is supposed to mean that separate observers reached the same conclusion through independent examination, while repetition is merely the same conclusion traveling through several mouths. The internet has become exceptionally good at making the second resemble the first.

None of this requires a conspiracy. It merely requires enough people following the rules long enough for the rules to begin impersonating reality.

This cycle depends on a peculiar asymmetry. Wikipedia demands identifiable institutional authority from the material entering the encyclopedia, while demanding almost none from the people deciding what enters, what stays, and what receives pride of place.

But don’t ever think Wikipedia lacks standards. It maintains elaborate rankings of publications it considers reliable, unreliable, deprecated, marginal, partisan, or otherwise beneath its dignity. These judgments are compiled, again, by anons, some of whom clearly enjoy larping as National Enquirer Employee of the Month, meticulously reproducing the same editorial instincts, without even the excuse of existential competition. Of the hundred possible ways to tell a story, pick the ugliest, then defend it as neutral on the grounds that someone else (whom you also chose) shares your taste.

Wikipedia, to its credit, advises that Wikipedia itself is not a reliable source. Yet this has not prevented it from honoring outside references whose wording, framing, and sometimes entire premise appears to have been laundered through its own unreliable content.

The deeper problem is determining what a “reliable source” even is. It is not a sterile unit of truth descending from heaven in PDF form. Publications have owners, advertisers, audiences, political cultures, class assumptions, commercial incentives, professional rivalries, institutional loyalties, access dependencies, blind spots, and grudges. They decide whom to cover, what to emphasize, which accusations deserve a headline, and which corrections receive six apologetic lines somewhere near the crossword. Wikipedia then imports those decisions, layers its own interpretation on top, and announces that judgment has been removed from the process. It has not.

Studies of English Wikipedia’s citation patterns have found systematic tendencies in news-source selection even after accounting for factual reliability. The precise direction matters less here than the broader point: selecting sources is itself an editorial act.

“We follow the sources” is not an escape from judgment. Choosing which sources count is judgment. Deciding which claims deserve prominence is judgment. Selecting one description from several accurate possibilities is judgment. Putting an insult in the opening sentence is judgment. Burying the objection ten thousand words down is judgment. Choosing the ugliest available photograph is judgment (this one hits close to home). Deciding that one participant’s departure constitutes consensus is judgment. Declaring that an article is neutral is judgment.

Even the prohibition against “false balance”, reasonable in obvious cases, does not remove judgment from the process. [WP:FALSEBALANCE: “While it is important to account for all significant viewpoints on any topic, Wikipedia policy does not state or imply that every minority view, fringe theory, or extraordinary claim needs to be presented along with commonly accepted mainstream scholarship as if they were of equal validity.”] Someone must decide when a controversy is genuine, when it has been settled, which dissent is substantive, which is marginal, and at what point acknowledging disagreement would become irresponsible. Those decisions may be justified, but they are still decisions. Giving an absurd claim equal billing is not neutrality, but neither is invoking "false balance" as a magic phrase that converts an editor's preferred boundary of permissible disagreement into an objective feature of reality.

Wikipedia has not eliminated editorial authority. It has disguised it as procedure.

Its editors are not disembodied instruments through which the sources speak. They are people selecting, weighing, framing, excluding, interpreting, and arranging information. The robes are templates. The gavels are warning notices. The courtroom is a talk page. Court stenographer apparently on indefinite wikibreak. The judge and jury are self-appointed, and the public is strongly discouraged from asking their names.

An encyclopedia written by human beings cannot be neutral in the absolute sense Wikipedia’s branding implies. It can certainly aspire to fairness. It can require attribution. It can display competing accounts. It can disclose uncertainty and discipline obvious prejudice. What it cannot do is evacuate perspective from the human mind.

“Objectivity” is not “contemplation without interest”… there is only a perspective seeing, only a perspective knowing.

—Friedrich Nietzsche, On the Genealogy of Morals, III, §12

His point was not that evidence is meaningless or that every interpretation is equally valid. It was that there is no view from nowhere, no immaculate observer floating above interest, culture, temperament, and experience. Wikipedia promises precisely such an observer, then places a pseudonym behind the curtain.

A credible neutrality policy would not require pretending that editors possess no perspective. It would require arranging the evidence so that readers retain some freedom to reach their own conclusions. A useful test is not whether an article has achieved the impossible view from nowhere, but whether the editors’ preferred conclusion can be detected in its architecture: in what is asserted as fact, what is attributed, what is explained, what is omitted, and what is buried.

Neutrality should restrain the editor’s judgment, not conceal it. The honest policy would not promise a Neutral Point of View, as though human beings could become surveillance cameras. It would promise something more achievable — Disciplined, Attributed, and Disclosed Points of View, evidence arranged so the reader can identify the judgment, locate its source, and decide whether to accept it.

But that would require admitting that the machine has operators. And Wikipedia’s favorite operator is the one nobody can see.

Rich, isn’t it. An institution assigning reliability grades to the rest of the information economy while being governed by individuals whose names, qualifications, employers, affiliations, and possible conflicts of interest are themselves uncited.

Transparency for thee, a burner account for me.

Share

In Part III: How a stranger’s exhaustion becomes “consensus”, and why silence is recorded as agreement.

This is a series about Wikipedia’s anonymous editorial power - and what happens when AI turns contested human judgments into facts.

Share

I regret making dinner plans. I’m over half an hour late, and my friend is now drinking alone at the Temple Court bar because instead of getting ready, I’ve been arguing with a Wikipedia editor named User:RedNipples* [redacted]. I have only slightly changed the color of the nipples. The user in question was one of the few who actually ended up being helpful, and my goal here is not to expose him specifically. With a username like that, my wild guess is that he is a man (like ~85-90% of their active editors, depending on which survey you believe), but of course, that is the point - I have no idea. 

Like an idiot, I tried following Wikipedia’s Byzantine policies. Since I’ve summoned enough audacity to request changes to my own page (a page that I wish didn’t exist in the first place), under my real name, I can’t be trusted to edit it directly, my only remaining recourse being politely pointing out inaccuracies (degrading insults, if you ask me), while maintaining the temperament of a hostage negotiator, like a beggar hoping to catch the eye of some merciful wikiwarrior who might stumble upon my pleas. I’ve been promptly blocked, unblocked, verified, accused of using an LLM because obviously it’s that hard to believe I’m capable of composing a paragraph about my own moral integrity, ignored, dismissed, accused of using the wrong font, but in the interest of keeping this piece under ten thousand words, I’ll save the full account of my personal Wikiwar in more juicy detail for a separate post. To sum it up: looking at my Wikipedia page really makes me wish I had opened a bakery in Eschweiler a decade ago.

Compared to the Wikipedia kangaroo court, jail was the lesser punishment - at least you sort of know the people, and the sentence has an end date. 

Like many people who have the misfortune of becoming the subject of their own WP:BLP (“biography of a living person”, duh! I could hardly resist showing off my freshly acquired command of all the priestly abbreviations), I found it easier to pretend mine didn’t exist. 

Would the same type of person who refuses to look at the page that unfailingly appears on top of every Google search about her also, on occasion, go to war over a minor inaccuracy or misspelling in some paywalled village publication, berating some poor sucker for lazy fact-checking and poor work ethic?

Oh, yes.

Until recently, I liked to think of Wikipedia as a relic of the early internet—dated, insular, a niche annoyance that mostly mattered to the unlucky people who managed to incur the wrath of someone identifying as a “wikipedian” and somehow ended up being branded with a biography, or WWE fans. An institution that somehow never closed: still standing, still influential in theory, but easy to assume the world had moved on without. It was easy to laugh off and dismiss. But that was before AI turned it into one of the world’s primary reference points.

When I came across the news of the encyclopedia’s cofounder Larry Sanger’s permanent ban, I went down the rabbit hole of the Wikipedia industrial complex and started to understand the scale of what an anonymous User:RedNipples has the power to affect. (For the sake of brevity, let’s use RedNipples as our token anonymous editor from here on.)

And suddenly, even as someone possessing advanced skill at living in denial, I felt rather extremely inadequate for deleting hours of my life misguidedly harping on inaccuracies in articles nobody could read past the paywall while looking away from the source code. That one page I refused to look at now shapes my Google results, my social media descriptions, my news coverage, my knowledge panels, my podcast introductions, and of course the answers produced by AI. Ask a chatbot who I am, and what you’ll get is in large part just Wikipedia recited back in a different font. 

The rules may not have changed much from the crowdsourced epistemic experiment’s perspective, but the world around them certainly has. What may have worked for a volunteer encyclopedia in 2001 no longer works for a major layer of global information infrastructure in 2026.

In the time I spent busying myself with untagging fugly pictures from my Instagram, Wikipedia became the master copy of consensus reality, while retaining governance rules designed for a hobbyist website from the early aughts.

Wikipedia still likes to peddle the democratic fairy tale of "the encyclopedia anyone can edit." They certainly would like you to believe that if so desired, just any rube can swoop in and provide his two cents, that his opinion will matter, and that his account will survive more than three edit attempts without being permanently blocked if RedNipples happens to be feeling sore.

And of course, in theory, anyone can *try* to edit almost anything — all you need to create a Wikipedia account is an email address. Just knowing the option exists should be enough for most people, right? Don’t mind the very high chance the rube’s edit will be reverted, but hey, wasn’t it fun to cosplay as one of the big boys of the basement Politburo? In the end, it’s about the friends we make along the way.

The scale of this oligarchy is not a secret (even though I’ve yet to meet anyone who cares as much as I do) - a Purdue study that crawled all 250 million edits from Wikipedia’s first decade found that roughly 1% of editors generated 77% of the site’s content- the lead author describes these editors as driven by a sense of mission “much like political or religious movements”. The English Wikipedia currently counts some 53 million registered accounts. Around 250,000 of them edited anything at all in the past month, and only about five thousand make more than a hundred edits a month. Of the three-quarters of a million registered editors who made at least one edit in 2024, about half were new accounts making their first edit — and, frequently, their last. Sixty percent of registered users never edit again after their first twenty-four hours. It must be the taste of freedom that makes them run for the hills, or maybe a forcefully blocked IP address.

For a site that demands a footnote and a source for everything, the RedNipples of the world certainly go easy on themselves. Editors are required to disclose nothing at all. No great shock then that accountability hovers somewhere near 4chan’s.

“Without the disclosure of the agent in the act, action loses its specific character and becomes one form of achievement among others… Action without a name, a ‘who’ attached to it, is meaningless.”

— Hannah Arendt, The Human Condition

Having browsed through quite a few editors’ pages (purely for research purposes, of course), I was unprepared for all the cheery goofiness and ill-advised attempts at comedy. Wikipedia has always struck me as a rather somber place. Its custodians, apparently, disagree. Some of their pages read like private standup practice - bits, riffs, entire routines - auditions for SNL held in their parents’ living room.

Note my restraint: I’m naming no names — not for lack of material, but because I’m doing my best to keep this installment friendly and neutral (WP:NPOV, hello), saving the acrimony for the follow-up.

Hate to make this about myself, but: I never asked for this atrocious page to exist. I would be perfectly content for it not to. I cannot delete it, opt out of it, or require the people maintaining it to identify themselves. It stands as a permanent monument to choices I’d rank among neither my greatest achievements nor my most interesting mistakes. The nearest analogy I can think of is the wrist tattoo I got at 19 in Berlin, except the comparison flatters Wikipedia, because the tattoo I can laser off whenever I get around to it. The page is there until the internet isn’t.

It belongs neither to me nor to any publicly accountable editorial institution. It belongs, in practice, to whoever is sufficiently persistent, procedurally fluent, and terminally available to prevail. And god forbid any of them be on my payroll — Wikipedia reserves its greatest suspicion for the one form of influence that comes with a paper trail (WP:COI).

And now, the closest thing I have to due process is the mercy of RedNipples.

Subscribe

In Part II: Why “reliable sources” aren’t, and how Wikipedia cites its own echo as independent proof.

Read Part II Here

Being in prison mostly feels like extended quarantining, only with a bunch of murderers, and we still can get our hair done.

I remember vividly the third Monday of March 2020, when Covid-19 first began its reign of terror. And just like that, in one afternoon, all visits, non-essential employees and civilians were indefinitely barred from entering the premises. I will never forget the scary moments when everyone thought the incoming packages and commissary were the next in line to be suspended. It will forever stay in my memory how, due to some interruption in the food supply chain, they took the liberty of replacing vegan French dressing with a regular blue cheese one. Like having to settle for non-organic mixed greens and broccoli wasn't punishment enough. 

I understand that prison is supposed to be all about having no rights and being forced to settle for less, but when I'm being casually told that granola bars are being replaced with powdered donuts, it's extra hard to maintain hope and stay positive. Anyways, I always thought that 'staying positive' is for morons who prefer to live in denial.

I survived the hours of uncertainty when they floated the possibility that cocoa butter may never be getting delivered to our commissary again. It wouldn't be fair to people with shea butter allergy/aversion, pretty much the only alternative available. In fact, it would almost amount to discrimination, and we all know what happens to those who violate local inclusion and diversity guidelines. They are never to be seen again. So, cocoa butter is back in the game.

Now that everyone's been getting used to their lives not going according to plan, mine is just right on schedule. Following the pandemic-induced elimination of all of the institutional brainwashing, my prison activities have been reduced to twelve-hour naps and nine yoga classes a week. I finally read Euclid's Geometry, Ulysses and Nietzsche's Thoughts Out of Season, mainly because I'm not a quitter, and those are the least popular titles from the complimentary eBook catalogue on our JPay media store. I wanted to see how much I can boost their rankings, but also to be able to say that I endured and wasn't impressed.

I made my own nut mylks, grew my herb garden, and tested every organic produce supplier that would deliver here. And if I'm not a better person now than I used to be couple months ago, I definitely am a thinner, and an inch and a half taller one. Can someone google and let me know if it's permanent or will I have to continue to stretch for two hours every day to maintain?

◆◆◆

At first when I started noticing #corona2020 tags in bathroom stalls in the gym, and seeing people getting it tattooed in the yard, I thought, 'what a romantic way to capture a moment in time.' The onset of the pandemic was not much different from most cultural shifts—the buildup was gradual, and then it all unraveled very quickly.

One day, corona was just another thing for the free world to handle, and the next day my yoga teacher is gone for good, never to be seen again. I never got a chance to say goodbye, or to let her know how much her breathing technique changed my life. She will never find out that I didn't really mean it when I said I couldn't relate to her when she wore that bright magenta Gap sweater. N, if you're reading this—come back? This is not a good time to stop believing that all those volunteer hours will eventually pay off.

Because now we're left to our own devices, which means we inherited a collection of DVDs, resistance bands and foam rollers, and it's just not the same. Two girls who used to be best friends-slash-co-defendants fell out over a dispute about Tracy Anderson vs. Mandy Ingber, and it turned incredibly ugly. I'll spare you the details, but let's just say that three veneers never left the room again that day. They haven't spoken to each other ever since. Over here, these things are not a laughing matter.

The same day, in order to adequately communicate the feeling of 'we-are-in-this-together'-ness, one of the staff makes an attempt at a joke—'If anyone here has cabbage and corned beef, I want it. They had none in the store. I will make you a good offer.' 

Cabbage & corned beef. Corned beef & cabbage.

'It's good, try it.'

Why, so I can be more like you?

◆◆◆

'You bougie.'

An omnipresent word that stands for an arbitrary concept by now devoid of any meaning.

It doesn't take much to reach bougie status.

I can see how managing to get in a 'nonflammable' Hermès blanket after twenty-two failed attempts could qualify one for the title, but how bougie is one really for declining a random head offer in a filthy yard bathroom stall?

Same goes for refusing to surrender the last thing you have control over—your body. Simply choosing not to eat two boxes of zebra cakes a day is already enough to join the bougie ranks.

Meet Debra—even though she is consistent with her routine, that fifth honey bun she religiously microwaves for four humiliating seconds at 10:57pm every night has so far failed to provide her with any supplemental guidance towards the path of enlightenment, but she's read something about perseverance in one of the motivational brochures and she keeps trying. That's right Debra, don't you ever give up.

'I'd offer you one, but you bougie', she'd say if I happen to witness it, not really meaning it, the same way people here offer me pepperoni pizza bagels and turkey fried rice, knowing well beforehand I'll decline, acting like they just remembered I don't eat meat, right after their generous offering.

◆◆◆

On the advice of sister Carmen, I drop a tab to try a popular local form of escapism—surrendering all rational thought and joining a religious group to find out if blaming the supernatural powers for my convictions is for me. Added bonus: a certificate from Jesus is supposedly 'good for the board.'

Board as in parole board, which I never get to find out because two days later my connect delivers the bad news:

'Our leader asked me to tell you that since we do none of the bougie stuff, she thinks you will disrupt the vibe of the group in a counterproductive way that's detrimental to the healing of other peace-seekers. Maybe next time. And also—“horrible penmanship.”'

'I'm trying, thanks.'

'She said she'll keep you in her prayers.'

'Tell Jesus I said hi.'

◆◆◆

After the initial excitement wore off, it became evident that the pandemic definitely took a toll on staff's morale. A good example is one of our steady COs from that period—now every morning he sits out on the front porch for five cigarettes, talking about the times way before he condemned himself to the life of polyester uniforms and mediocrity, watching a local family of deer watching him through the diamond wire fence.

For years he'd been blind to the madness of this place, and now it's all catching up with him.

I can tell he is in contemplative mood, and he is sad. He is sad because no one told him that he won't be reforming dangerous criminals but instead adult-sitting a bunch of drug addicts whose fights over their girlfriends are the most exciting and wild thing that will ever happen to his career, so I can't blame him for drinking on the job.

It's been over a decade since 'Care, Custody and Control' were first forced upon him as his core guiding principles, and he's wondering what life would've been like had he chosen that job at Walmart across the street from here like his grandma told him to. So close, and yet so far. Just out of reach.

With a jittery hand he vacantly fumbles at the torn S on his reused SlimFast bottle, wondering if it's fooling anyone other than himself, and for a moment he looks so helpless and lost, it almost makes me want to hug him. Too bad I'm not a hugger. And whoever said I had no feelings.

◆◆◆

The mighty administration of this correctional facility informs us about their readiness to rule on some of the more pressing issues at hand.

It's very curious to see an institution so inherently averse to change scrambling to adapt on the go; this is when the mysteries of higher decision-making come to light and the real power players are being exposed.

First comes a newsflash about the novel hand sanitizer. They send the most likable ones to deliver the most devastating news. Sergeant A is a big hit with the local population. He simply can do no wrong, and is always relied upon to perform the hardest tasks. He rarely disappoints.

'Alrighty, girls, since last night there have already been two cases where some of you decided it was a good idea to drink the hand sanitizer. It still remains to be seen if there are any survivors. I never thought I'd be saying this, but they sent me to make a round and advise you against drinking it. Clearly, we cannot put it past any of you any more. You don't want to find out firsthand how inadequate the medical staff here really is. Just think about how if you mix yourself a cocktail made out of state supplies you won't even get a chance to collect that lawsuit money for those mistakenly pulled teeth of yours from '97. To be honest, I don't care either way, but if you drink it, you'll probably die.'

Nothing like a father figure with unshakable belief in one's potential.

He reminds us that to help staying focused, there is an awesome selection of educational videos and suicide prevention movies available for us at all times to view on our tablets.

For the sake of research, of course, I have to check those out.

And there is something for everybody. Alternatives to violence, building mindset for success, anger and addiction management, Christian stories, childhood trauma, general self-improvement.

I mean, how can I possibly be expected to choose here. Jesus, can someone just tell me what to watch already?

◆◆◆

I decide to watch them all. Why not?

Seeing all these overachievers on one screen also makes me think how I'm totally not reaching my full inmate potential. I could do so much better, only if I tried a little harder.

The one episode that instantly jumps out is about Andre who learns how to cope with his 15-year sentence for a drug sale by learning to look at it as a blessing.

The video wants to help me see how 'it's possible to turn your time away into ultimate source of pride for your family and yourself.' It's an actual quote I got from the description. Are they, like, advertising prison?

I've heard it's working—is that why so many are coming back? Or maybe it's because they just can't wait to get back to their janitor jobs wearing new Jordans and reeking of counterfeit Gucci Guilty.

Too bad the cinematographers forgot to add 'For some more prison propaganda, follow us on Instagram and Twitter' and 'Getting Out With Grace coming to a correctional facility near you.'

It gets better.

For over twenty minutes, Eric is talking about the importance of taking re-entry to society seriously, while Joshua is explaining how he chose to move on and forgive himself for committing a murder in first degree and advises everyone to follow his lead. Basically the usual aspirational/motivational stuff.

My personal favorite is the 'Do It (or Don't)' episode. It's not always this easy, but it's often this easy.

I skip 'You've got a choice' and 'It's all in the family,' since incest is not a topic I wish to explore further. Finally, I find something that resonates with me—'Getting Past Addiction,' which reminds me of that time I went to rehab and beat my sugar addiction. Sugar is hundred times more addictive than cocaine and is number one cause of craving more sugar, which leads to premature aging and death. I am a survivor.

Hearing them discuss video games, social media and sex addictions, I am comforted by the fact that the creators genuinely seem to be convinced that every addiction is interesting and important, which is what I'm looking for in a video like this—the feeling of shared adversity.

I stop watching because there is a common thread I notice that bothers me—all these videos, without exception, are based on male characters. It's like female-committed crimes are somehow inferior and therefore deserve less attention. Like they don't trust women to be capable of an offense that's tragic enough to be worthy of a stand-alone episode of their shitty little series. A three-day all-access pass to my facility would provide them with at least five seasons worth of inspirational material.

But no. In their narrative, the female, instead of focusing on pursuing her own crimes, is always introduced as a blissfully static presence busy trying to point the prodigal son in the right direction. No, motherfucker, what if it's Miss Gracie's turn to get crazy? 

Just joking.

Whatever.

◆◆◆

While I'm busy building my mindset for success, a fire alarm goes off, and this time it's inside our bungalow. Now the Fire Safety Sergeant is being summoned. Since that's his only job title, this incident presents a rare opportunity for him to shine and show off his impressive skillset. Sporting a cap and a mustache, he comes through looking like Super Mario. My neighbor, D, can't resist: 

'Are we going to make it? I refuse to die fat.'

'It doesn't look good, but I'll do my best,' he responds. 

Ten times out of ten the fire alarm gets triggered as a result of someone's cooking, and this time it’s Mamacita's banana bread. Under Super Mario's guidance the offending carb bomb gets neutralized swiftly and everyone resumes their criminal activities.

We are so incredibly lucky to have a dedicated team of professionals keeping us safe and alive during these uncertain times. 

Using this forced intermission as an excuse to get everyone in the same place, we proceed smoothly into a 'community meeting.' Those are always made up of the same trivial stuff—quarantine updates, new tennis tournament schedule, moderating Amish Mafia vs. Big Shirley chore wars, cube compliance, think before you sue, we never get enough ice, bingo nights, boring.

Our fearless leader, the community-elected ILC representative (if at any time during your incarceration you get in touch with your power-hungry side, Inmate Liaison Committee is a hobby for you), informs us of her executive decision to order new, next-generation sanitizing supplies and copper cooking ware. That's for others since I don't cook or clean, so whatever.

The total of three people seem to be paying attention, the rest are immersed into the all-consuming Fish Master, 'the wet and wild fishing game,' with their headphones on blasting what sounds like Viral Moment by Lil Durk. Rude. Love them.

Then she comes through with the ultimate highlight in form of a round of free puzzle books for everybody, sponsored by our bottomless inmate fund. Nothing like some connect-the-dots and crosswords to take one's mind off things and unwind after a stressful day. I can see how after recovering from a forty-year-long freebase addiction, a sudoku is the first thing one automatically reaches for in order to deal with the pandemic-induced anxiety.

My other neighbor, V (the one that tells everyone she fucked Banksy), is getting restless. She is determined to catch the 2pm zumba class before her competitive knitting session, and this is cutting into her schedule. 

Not so fast, the fun is not over, not just yet.

We are awaiting for the higher powers of this reform school for wayward girls to descend upon us for a weekly a convention to fill us in on the current affairs. Just joking—they come primarily to sign 'the Book' and flee as soon as they can.

By this time, the suspense is killing me and everybody else in the bungalow. I just finished both the latest Vogue Italia and 032c I've been reading simultaneously and it seems like inappropriate time to start anything new, much less to get into the new CR Fashion Book. Don't you just hate being interrupted in the middle of Carine Roitfeld-orchestrated cover spread? The worst. I'll wait.

As a premonition of all the tired cliches coming our way, the two that show up look like a walking number 10—one slim and tall, the other round and short. The good cop and the bad cop. The latter squinting in his ubiquitous wraparound mirrored sunglasses, potentially channeling the South Park Satan, philistinism dripping off of him.

They roll into our den of iniquity, determined to have it done and over with quickly. Their group order from the local upscale-casual Korean BBQ, one of the few places that are still open and solvent up here, will arrive at the front gate in less than 15 min. I know this because it's been announced over walkie-talkie for all the 'radio holders,' and these days I'm never too far away from one. In certain places, like the gym or the yoga floor in the school building, the officer-to-inmate ratio often nears 3-to-1. A typical Saturday night is 9 COs playing poker on the back porch while 3 inmates are doing yoga inside. Security is no joke.

The good cop takes the lead with an expected mandatory 'There are no confirmed cases anywhere on this compound' and 'We are praying (paying?) for everyone's safety,' (oh please, as if), followed by 'If due to visiting cancellation your drug supply has been cut short and you're detoxing and/or experiencing withdrawal symptoms such as hopelessness and depression, please don't kill yourself just yet. We understand that dealing with sobriety can seem like a daunting and insurmountable task even under regular circumstances, and more so now, during the times of pandemic. That's why we have some great news for ya'll. Starting yesterday, you guys have access to select free movie and game titles available for download on your tablet. In addition to that, you get some free phone calls to connect with those loved ones who otherwise wouldn't put money on their phone accounts. Have fun!'

The bad one then adds 'Enjoy!'

Note the masterful application of the classic 'deny and deflect' maneuver, a proven way to pacify discontent. You know the dirty feeling when you've been advertised to and it worked. The mood now is something like that.

The short fat one stares at us, like he just blew his own mind with this burst of generosity. Three people stare blankly back at him. One of them doesn't even live on this dorm, having been stuck here for the past hour visiting her gf.

'Ok, ladies, have a good night and drive home safely.' Complete smugness.

'And, oh—if you don't like it here, leave.' Done, out. The End.

◆◆◆

Bad news is that the free games weren't Aporkalypse or UberClaus, as everyone hoped they would be. Maybe next month. Instead they donated some unmemorable demo versions of something with bouncy balls and bricks and arrows, apps that looked like they were a talentless middle-schooler's uninspired side project. Felons can't be choosers, I guess.

At least the free movies weren't home-made, but sober Lindsay Lohan in Labor Pains just isn't the same.

And I'm never getting those 2.5 min back I spent scrolling forward in Sharknado. Why did they ever make this? Later we discover that all seven parts will eventually be bestowed upon us, one-by-one on monthly basis, clearly so we have things to look forward to during these hopeless times.

The rest was something marine-themed and one vanilla Christian feel-good movie. I'll never find out what those were about.

Scrolling through the 'Editor's Picks' of the media store I realized that during those past months I finally learned to accept and made peace with the fact that now is not a good time for me to use a red pen (those are for higher powers only), wear blue shoes, or eat seaweed snacks. I start to see how an outsider can be led to believe that I may not be prepared to handle these things, reserved exclusively for the people on the outside, not yet. But maybe sometime very very soon, when the higher powers see all the great strides towards self-improvement I took while assembling a whole new monochrome wardrobe, fearlessly navigating countless hurdles, and realize that I'm so ready.

So while I'm wearing 100% cotton panties, eating leaves and looking for the meaning of this invaluable state-mandated experience, I decided to create an absolute non-communication, a perfect self-sufficiency. A retreat within a retreat of sorts. 

Let's hope I find it before I develop a Stockholm syndrome, become religious, or worse, fat.

— Anna Delvey
Albion, NY
2020

Voyeuristic headlines, equal parts degrading and incendiary, herd the algorithmically vulnerable, gullible and easily clickbaited into a regurgitation of the same quota-filling disquisition. 

Here is a step-by-step guide:

Procure the subject’s third- and fourth-ugliest photographs available. (Save the first and second for special occasions. The future is bright.)

Repeat every negative fact everyone already knows until it occupies at least three-quarters of the article.

Ignore any positive development or redeeming quality.

Press publish.

Congratulations. You’re now a Daily Mail reporter. (Sorry, Daily Mail, nothing personal.)

No masterpiece is complete without the obligatory mention of “fake heiress.” Years ago, everyone collectively decided it struck the perfect balance between condescendingly amusing and sufficiently humiliating, and it’s been everyone’s absolute favorite go-to ever since. It’s often followed by “con artist,” just in case anyone needed clarification, and then some combination of “criminal,”, “convicted felon”, “illegal alien,” “deportation,” and “hopefully to Russia” to really drive the point home while also maximizing SEO performance. The excitement of public punishment disguised as civic concern.

This is inevitably followed by a round of outrage porn detailing my greatest criminal hits, carefully enumerating charges from nearly a decade ago, should anyone dare forget them. And of course each piece is accompanied by a thoughtfully selected assortment of outdated photographs, preferably the ones featuring handcuffs or prison uniforms, providing the jerk-off material for readers yearning for a safer country and liberated taxpayer dollars that were squandered on my ankle monitor. The same one that I love to bedazzle and dance in and photograph for my Instagram and then turn a profit on. The auteurs of these pieces are acutely aware that without the ankle monitor, my entire identity would be erased, reverting me to my factory settings: regular, plain, unadorned, unsupervised, irrelevant. Just like everyone else. 

Deserving special mention is an offense so shameless and contemptible, it’s remarkable I keep going: posting on social media while wearing an ankle monitor.

An act so sick and sinister, so scandalous and so daring, that it alone has become the subject of multiple think pieces. No amount of belittlement or disbelieving head-shaking ever seems sufficient. The mere fact of my existence is an act of reckless defiance, some grand affront to fragile sensibilities.

Few might know this fun fact, but paths towards your own ankle monitor are not one but plenty, and not all of them involve an unwavering commitment to a life of crime. Different models help authorities with different tasks. Some are meant to police alcohol consumption after a DUI. Others are used during criminal pretrial supervision when there’s a risk of absconding. Some enforce curfews during parole. Others are used during immigration proceedings to ensure the wearer doesn’t disappear into the country and live happily ever after as an undocumented fugitive.

To the untrained eye, however, the device itself—an updated scarlet letter in the form of a black plastic box attached to a rubber strap—looks exactly the same regardless of purpose. Absent context, the wearer could be anyone from a serial killer to a foreign nun with visa problems.

Once you’ve completed the hard part—executing the choices that led to your current circumstances—you are presented with two options. You can cover it up, or you can leave it visible.

Both choices are wrong.

You’re either flaunting it or hiding it. You’re either insufficiently remorseful or performing remorse manipulatively. Reprehensible either way. 

***

The best thing about being publicly hated is that eventually you realize most people aren’t actually thinking about you very much. They consume scandal the way they eat DoorDashed French fries on the couch watching Netflix while scrolling TikTok after a long day: mechanically, absentmindedly, without much pleasure and hopefully without any long-term memory.

I also like telling myself that most people don’t even particularly hate you. They simply like to enjoy that brief feeling of moral superiority. For thirty seconds, while reading about someone else’s shame, they get to feel a little better about themselves. Innocent. Superior. Then they think about what they’re going to have for dessert.

Redemption is wonderful in theory. But in practice, redemption is apparently just deeply offensive, and surviving humiliating circumstances a little too well becomes its own kind of crime.

The public says it wants accountability, but what it actually wants is permanence. Permanent guilt, permanent silence and permanent shame. An identity frozen in its most unflattering moment and preserved until the end of time in a quick Google search result. A life sentence of public humiliation, after which you’re expected to die as the version they decided on.

Journalism as taxidermy.

Punishment must also be visual and easily consumable, and it’s only satisfying when it appears endless, anything less feels incomplete. The imagery ready to get eaten up - mugshots, surveillance footage, public collapse, reaction videos, all preferably clipped under one minute long and accompanied by a trending sound.

Visible suffering reassures everyone that the machine is functioning properly. In order to be believed, there needs to be tangible proof of ruin.

Ideally, justice would amount to total obliteration. But if a comeback story is what’s on the menu, it’s only digestible if the person returns humble enough to flatter the audience’s sense of moral superiority. It also wouldn’t hurt to spend the remainder of your life performing gratitude for being allowed to remain alive. 

What’s fascinating is how angry people become when the person they reduced to a cautionary tale refuses to remain flattened by it. God forbid anyone thinks you adapted instead of collapsed. What adaptation actually looks like is far less cinematic, and rarely gets broadcast—the same way you hope your friend doesn’t post the pre-FaceTuned version of the photo already sitting in your carousel. 

Refusing to collapse into shame is not the same thing as refusing accountability. Nor is it glorifying crime, or making mockery of the system as a whole. But that distinction isn’t particularly conducive to a salacious narrative, so it rarely matters. 

“If I say ‘I forgive you on the condition that, asking forgiveness, you would thus have changed and would no longer be the same’, do I forgive? What do I forgive? And whom? … Something or someone? … Does one forgive something, a crime, a fault, a wrong … or does one forgive someone, absolutely, no longer marking the limit between the injury, the moment of the fault, and on the other side the person taken as responsible or culpable?” (Jacques Derrida, On Cosmopolitanism and Forgiveness, p.38)

Another grievance thrown my way is the alleged “profiteering from my crimes.” One would think repaying restitution, fees, and fines would render that point moot, but that would be letting me off too easily.

Apparently, I’ve managed to profit from jail, prison, court appearances, house arrest, ankle monitor - profit simply follows me everywhere I go!

It taught me compassion—constantly having my hard work dismissed as a mere byproduct of my criminal past has made me unexpectedly sympathetic to the plight of everyone ever accused of benefiting from nepotism.

Meanwhile, people have built entire content libraries, and in some cases entire identities, analyzing, monetizing, condemning, defending, dissecting and diagnosing me.

Everyone is free to profit from my crime except me. No fair.

Apparently, my story is best told by an upstanding citizen who is not currently wearing an ankle bracelet.

I’d love to argue that most of what I’ve managed to accomplish has happened in spite of my past, not because of it. Life is already hard enough under normal circumstances, and even though I’m all for sprinkling in a healthy amount of adversity here and there, adding years of legal obstacles rarely improves the experience.

Who gets to decide when the illicit profit ends and the legitimate one begins?

At what point, exactly, does someone stop being defined exclusively by their worst decision? Is it immediately following release from custody, the way we restore someone’s right to vote and possess money but not necessarily their right to possess dignity? Or maybe at some point in the future, like your right to own a firearm? Or is the answer Never? Because saying “well, you shouldn’t have committed a crime in the first place” means that punishment doesn’t end when the sentence does.

It simply changes form.

Is there a parole board for public opinion?