As a writing instructor, I have spent years arguing against the kinds of assessments they represent. Timed, handwritten exams have always felt disconnected from the way people actually write. Professional writing is iterative. We brainstorm, draft, revise, seek feedback, and rethink our ideas. In compositions studies we say it incessantly: writing is a social process. No one sits in a silent room with a pen and a blank booklet to compose an important email, proposal, or report.

So when generative AI arrived in higher education and instructors began calling for a return to Blue Books, I wasn’t convinced. It felt like a reaction driven more by fear than by pedagogy. Trading laptops for lined paper seemed like trying to solve a twenty-first century problem with a nineteenth-century solution. And I still don’t think Blue Books are the answer.

In all honesty, I hesitate to say this conversation is even a debate between Blue Books and AI. The future of assessment is not going to be won by pens or by algorithms. Rather, it is a conversation around how we ask students to make their thinking visible—and the equitability around this “visible thinking.”

For the past two years, much of the conversation around assessment has revolved around academic integrity. How do we stop students from cheating? How do we prevent them from using ChatGPT? How do we prove that the work they submit is really their own?

These are understandable concerns, but I have found myself becoming less interested in them. Instead, AI has pushed me toward a different question altogether. What evidence of learning do I actually want to see?

For most of my teaching career, the answer seemed obvious. I wanted to evaluate the final product. I wanted to know whether students could write an effective business letter, organize an argument logically, or communicate professionally to a specific audience. The finished document was both the assignment and the evidence of learning. Evidently, AI has complicated that assumption.

Today, a polished document no longer tells me everything I want to know. Students have access to tools that can generate grammatically correct prose, reorganize paragraphs, adjust tone, and even imitate disciplinary conventions in a matter of seconds. Whether or not students choose to use those tools responsibly is an important conversation, but it has also exposed a weakness in many of our assessments. We have often treated the final product as though it were synonymous with learning, but I’m slowly becoming less and less convinced by this.

I know I’m late to this conversation.

Ever since ChatGPT burst onto the scene, educators have been urging one another to “make learning visible.” If students explain their reasoning, annotate their drafts, or defend their choices, AI becomes less useful because the thinking itself becomes part of the assessment. I agree with all of that, but I also think we’ve accidentally framed visible thinking as an anti-cheating strategy rather than a learning strategy.

If the primary reason we ask students to explain their thinking is to prove they didn’t use AI, haven’t we simply built another surveillance mechanism into our classrooms?

But if the reason we ask students to explain their thinking is because understanding their reasoning gives us better evidence of learning, then we’ve changed what the assessment is for. The assignment may look exactly the same, but the pedagogy is completely different.

Recently, I read a thoughtful reflection by Yvette Iribe Ramirez on the renewed interest in Blue Books in the age of generative AI. Ramirez argues that bringing back handwritten exams should not be about nostalgia or punishment. Instead, she suggests that analogue writing can create opportunities for engagement, reflection, and human connection that are increasingly difficult to cultivate in technology-saturated classrooms. She cautions against viewing Blue Books simply as an anti-AI strategy and instead frames them as one tool within a broader pedagogy of engagement.

That argument resonated with me, but it also helped me articulate something I had been struggling with. I’m not actually interested in bringing back Blue Books. I’m interested in bringing back thinking. And my whole plight against Blue Books may have been centered around the framing through which I was viewing using the Blue Books. I did see it as a punishment. A way of controlling my students through protectionist means that felt rooted in distrust rather than learning. They seemed to communicate, "I don't trust you with technology, so I'm taking it away." That framing never sat well with me because it positioned assessment as surveillance rather than an opportunity to understand what students know and can do.

Around the same time, I revisited a study examining the history of Blue Books in American higher education. What surprised me was how familiar the debates sounded. Faculty members appreciated Blue Books because they supported essay-based assessment and reduced certain forms of cheating, while students generally preferred typing and reported greater anxiety with handwritten exams. Even before ChatGPT existed, instructors were wrestling with many of the same questions we face today: What should writing assessments look like? How do we balance authenticity with practicality? Will new technologies eventually replace handwritten exams? The researchers concluded that although Blue Books would likely continue declining, they would probably never disappear entirely because they occupy a unique place in the assessment of student writing.

Many educators in the online space have been saying what these papers are insinuating. That the real issue is whether our assessments allow students to reveal their thinking. This semester, I’m planning to experiment with a different kind of Blue Book exam in one of my business writing courses.

Share

Students will write a professional letter responding to an employee, customer, or client. At first glance, it will look like a traditional handwritten assessment. The difference comes after the writing is finished.

Once students complete the letter, they will identify three places where they made deliberate rhetorical decisions. They will circle those moments and explain them on the following page. Why did they choose that opening? Why did they organize the information in that order? Why did they soften the tone before delivering difficult news? Which concepts from class shaped those decisions?

In other words, I don’t just want them to evaluate the letter. I want them to evaluate the thinking that produced the letter. To me, the metacognitive reflection is, in many ways, more important than the writing itself.

Educational research has consistently shown that metacognition—the ability to monitor and explain one’s own thinking—supports deeper learning. Similarly, research on retrieval practice demonstrates that students strengthen their understanding when they actively recall and apply knowledge rather than simply recognize information. Asking students to explain why they made particular rhetorical decisions draws on both of these principles. Instead of merely producing an answer, students must demonstrate that they understand the reasoning behind it.

That strikes me as a far richer form of assessment than simply asking whether they can produce polished prose.

Importantly, I also don’t think this idea depends on Blue Books.

The same approach could work through digital annotations, oral reflections, recorded think-alouds, or written commentaries attached to a draft. The handwritten format is simply one possible environment in which students can demonstrate their reasoning without relying on external tools.

Perhaps the most important lesson AI has taught me is that good assessment has never been about collecting finished products. It has always been about finding meaningful evidence of learning. For a long time, I assumed those two things were the same.

I’m no longer sure they are.

So yes, I’m changing my mind about Blue Books.

Kind of.

I’m still not interested in returning to an era of handwritten exams simply because they’re harder to cheat on. I still believe students should learn to write with digital tools, revise their work over time, and use AI ethically as part of an authentic writing process.

What has changed is my understand of what deserves our attention. If AI has given us anything, I hope it’s the opportunity to stop treating writing as a product to be graded and start treating it as thinking made visible.

—Dr. Sydney

**A side note because I’m genuinely still wrestling with this: One thing I haven’t fully resolved is whether an assessment that emerges from concerns about AI can ever truly escape those origins. Can something designed, at least in part, to limit AI use become good pedagogy? Or are we simply rebranding surveillance with more pedagogically appealing language? I don’t know the answer yet.

My instinct is that the difference can’t simply be intention. Every educator can say, “I’m doing this because it’s good pedagogy.” That isn’t enough. There has to be something more concrete that distinguishes an assessment designed to catch students from one designed to understand them.

That’s the question I’m still thinking through, and I’d genuinely love to hear your perspective. Where do you think that line is? When does an AI-resistant assessment become an authentic learning experience rather than just surveillance in disguise?

In media literacy scholarship, the protectionist approach was one of the earliest paradigms. Emerging in the mid-twentieth century alongside concerns about television, film, and mass media, it positioned audiences, especially children, as passive recipients who needed to be shielded from harmful messages. The assumption was pretty straightforward. The media were powerful, audiences were vulnerable, and education’s job was to protect them.

Then the field evolved. Media literacy shifted toward teaching analysis and evaluation rather than avoidance. Later, critical media literacy pushed even further, asking students not only to interpret media but to question who creates it, whose interests it serves, and how systems of power shape what we see. Instead of treating students as empty vessels waiting to be filled with the “right” answers, critical media literacy assumes they possess agency. They are capable of questioning, resisting, creating, and making informed decisions for themselves.

That belief has become foundational to my own teaching.

I’ve never wanted to rule my classroom with an iron fist. I’ve never believed education works through surveillance or punishment. The idea that we can—or should—protect people from every harmful technology starts to feel a little too close to Orwell’s 1984 for my comfort. If students never have opportunities to exercise judgment, are we actually teaching them to think critically?

And yet...The classroom has become the place where my philosophy collides with reality.

Can I really expect students to stop using AI simply because I teach them about its ethical, environmental, and cognitive implications?

Can I assign research on declining critical thinking, discuss the environmental costs of massive data centers, or facilitate conversations about authorship and learning, and expect students to respond, “You’re right, Dr. Sharkey. I’ll never use ChatGPT again.”

I’m not naive enough to think yes.

Leave a comment

Students are exhausted. They’re juggling work, caregiving responsibilities, anxiety, financial stress, and impossible schedules. Sometimes they’re sick. Sometimes they’re overwhelmed. Other times, maybe they are genuinely just being lazy. No matter the excuse—valid or invalid—convenience is powerful, and AI offers convenience in abundance.

Lately I’ve found myself thinking a lot about ’s newest book, Eat the Stars and her writing on climate grief. You’ll have to read it, it’s a good cry—both happy and sad.

She argues that hope, at least as we usually define it, can become its own kind of trap. We keep waiting for evidence that everything will be okay before we decide to act. But there comes a point when certainty disappears and there is a strange relief in accepting that we may not be able to fix everything. Oddly enough, I’ve begun wondering if AI in education demands a similar kind of thinking. Not hopelessness in the sense of giving up, but in the sense of relinquishing the illusion that I can control every outcome.

As a first-generation college student, education changed my life. It gave me confidence when I had very little. It taught me how to think, not just what to think. It gave me purpose. It made democracy feel personal because it taught me that ideas matter and that my voice mattered too.

That belief is still at the center of why I teach. But maybe my responsibility isn’t to save every student from AI.

Years ago, when I first started teaching, a mentor told me something I’ve thought about often: Focus on the good ones. At the time, I hated that advice. It sounded elitist and dismissive of students who struggle. I was one of those students. Plenty of brilliant people don’t arrive in college polished or confident or motivated every single semester. But over the years, I’ve tweaked that advice to: Focus on the ones who care.

There are still so many students who are still willing to wrestle with difficult questions. Students who are curious enough to revise, to reflect, to wonder whether outsourcing their thinking is actually costing them something.

I can’t police every suspicious em dash. I can’t investigate every oddly polished paragraph or every sentence that begins, “It’s not that we’re seeing this. It’s actually that.” I also can’t build an educational philosophy around suspicion.

However, I can certainly create classrooms where thinking is valued, where reflection feels worthwhile, and where students encounter ideas strong enough to make them pause before asking AI to do the thinking for them. Critical media literacy has always argued that education isn’t about controlling people. It’s about helping them develop the capacity to choose.

Maybe AI hasn’t changed that as much as I thought. It’s simply made the stakes of choice impossible to ignore.

-Dr. Sydney

Subscribe now

So, before I get to what they said, here’s the rundown of the pedagogical ideas we were working with…

Last week I assigned them “AI Killed the Take-Home Essay. COVID Killed Attendance. Now What?”

Here are some of the main pedagogical ideas from that piece (though please read it if you have time! First-hand accounts are always better than summaries):

• Mandatory attendance

• Daily in-class handwritten writing

• Shift away from take-home essays due to unverifiable authorship in the AI era.

• Assessment centered on observable work: discussions, presentations, facilitation.

• Students as co-creators, responsible for introducing readings and leading discussions.

• Public performance as learning, requiring real-time comprehension and synthesis.

• Structure and routine used intentionally to rebuild attention and engagement.

• Focus on authentic cognitive labor, not polished AI-assisted products.

• Classroom as an intellectual community, not a space for private, outsourced production.

• Messiness is embraced as a sign of genuine learning rather than failure.

• Confidence built through demonstrated competence, not polished prose.

• Liberal education reframed as civic work, cultivating skills essential for democracy.

• Call for institutional change, not just individual instructor adaptation.

To help students think a little more critically about this piece I also offered this scholarly article, “Teacher Influence in the Classroom: A Preliminary Investigation of Perceived Instructor Power, Credibility, and Student Satisfaction.”

This article outlined several types of instructor power:

• Expert power: demonstrating knowledge or expertise

• Referent power: building positive relationships and rapport

• Reward power: providing encouragement or incentives

• Coercive power: relying on threats, punishment, or strict control

I thought this would be a quick lesson in which all students saw the article as implementing only one type of power, my guess being coercive, and then we would move on from there. And wouldn’t you know it, I was sooo ironically wrong. The irony of this being that what I sought to do was prove that instructors don’t always know best. And neither do I!

The main ideas in AI Killed the Take-Home Essay. COVID Killed Attendance. Now What? have been circulating the internet. Much of which discusses strict rules such as attendance policies, students “teaching” the class, and no take-home assignments. The author makes a point of emphasizing the need to “witness” all of the assignment’s students produce, in order to validate their knowledge. All of which I saw falling into the category of “coercive power.” However, I was blind to many of the elements students also saw, such as student’s being “co-creators of the learning environment.”

To set up this class discussion I gave them a shared Google Doc with the following prompt:

We read “AI Killed the Take-Home Essay” and Teacher Influence in the Classroom: A Preliminary Investigation of Perceived Instructor Power, Credibility, and Student Satisfaction.

The scholarly article specifically discusses the differences between expert power (showing knowledge), referent power (building positive relationships), and reward power (providing encouragement or incentives) and coercive power (using threats, punishment, or strict control). Where do you think some of the ideas from AI Killed the Take-Home Essay. COVID Killed Attendance. Now What? fall into in terms of power? What makes you say this? Pull 1-2 quotes from the article that show which type of power it is using.

Then they got a table where they could fill out their group names, the power, and a quote. I’ll give you all a voting poll! So, what power would you label some of the pedagogical ideas from the beginning of this Substack?

Here is a summary of my students’ answers:

Referent power

Expert power

Referent Power

Coercive power

Referent Power

Expert Power

Coercive power

Referent Power

Reward Power

It’s really…all over the place. And as we discussed more in class, what became even more interesting to me was their reasoning behind why they labeled the power the way they did. One student remarked how they could see how the author was trying to use referent power, but it wasn’t landing that way in the writing. Other students discussed how shame and humiliation were big “motivators” for them in a class—and that’s why they could see coercive power being useful, despite the obvious negative wellbeing impacts.

Share

Throughout this activity I felt reminded—as I am constantly—of the enormous divide between the pedagogy we intend for our students and how our students actually experience our teaching. They are so rarely aligned. In some ways, this can be annoying. Can I really never create a pedagogical foundation without consulting my students on each and every concept or practice I use? Maybe not. But truly, why would you want to? When students are the core reason we teach, why would we want a disconnect between our intention and their experience? If anything, we should constantly be aiming to close this gap in our practice and pedagogy.

What this week ultimately clarified for me is that conversations about power in the classroom aren’t theoretical—they’re lived. Students read our choices, our policies, and even our tones in ways we can’t fully anticipate. And instead of seeing that unpredictability as a threat, I’m learning to treat it as an invitation: to listen more closely, to design more collaboratively, and to let my students’ interpretations shape the questions I ask next. If AI is transforming the nature of learning, then perhaps the most important skill we can model is not more control, but more curiosity.

One of the first discussion boards we’ll tackle next week is built around the post “AI Killed the Take-Home Essay. COVID Killed Attendance. Now What?” I stumbled on it during one of my weekly Substack scrolls and immediately flagged it. I’m always drawn to pieces like this—not necessarily because I agree with them, but because I’m curious about what students make of these sweeping claims about their generation, their work ethic, their relationship to technology, and their engagement with school.

Subscribe now

Here is our discussion board for next week:

AI Killed the Take-Home Essay

For next week, please read the Substack post “AI Killed the Take-Home Essay. COVID Killed Attendance. Now What?”

I’m especially interested in your perspectives on arguments like the ones made in this piece. Many instructors seem to approach AI from a position of control—framing it as a threat that needs to be contained rather than a tool that needs to be understood. This post leans into some of those themes. What do you think of the argument the author is making? Does it resonate with you? Does it feel off-base? Something in between?

To help us think more critically about this conversation, I’d also like you to read the scholarly article “Teacher Influence in the Classroom: A Preliminary Investigation of Perceived Instructor Power, Credibility, and Student Satisfaction.”

This article outlines several types of instructor power:

• Expert power: demonstrating knowledge or expertise

• Referent power: building positive relationships and rapport

• Reward power: providing encouragement or i

  ncentives

• Coercive power: relying on threats, punishment, or strict control

As you read the Substack piece, consider the author’s stance through this lens. Where do you see examples of these kinds of power showing up in their argument? What makes you say that?

Please post brief thoughts (about 150 words). We’ll save the deeper unpacking for class discussion.

We’re discussing this particular post—and intentionally linking it to a scholarly text on power—not because I’m leading students toward a predetermined conclusion or hoping they “figure something out” about the author’s argument. Instead, I’m genuinely curious about their reactions. When instructors talk about AI, we often default to big-picture narratives about cheating, accountability, or student motivation. But when students talk about AI, the tone is very different. They’re thinking about stress, survival, workload, digital exhaustion, access needs, inequity, and the feeling of being constantly assessed by instructors, by algorithms…really by everything.

So, when I say I’m interested in their perspectives on how power operates in these conversations, I mean it sincerely. Scholars have been writing for years about the “power of AI” in educational, social, and cultural contexts. And even longer about the power dynamics that underpin classroom relationships. But what we haven’t quite figured out—because AI is still so new, so fast-moving, and so entangled in the everyday realities of teaching—is how instructor power and student power shift in relation to AI use.

Does AI redistribute power? Inflate it? Complicate it? Blur it?
Who gets to decide what “ethical use” looks like?
What counts as learning now, and who gets to define that?

Mintz describes the current moment in education as a collapse, essentially an instructional ecosystem dissolving before our eyes. But I’m not convinced the ecosystem has collapsed so much as it has revealed itself. Students have long been navigating impossible workloads, inaccessible assignments, unreadable expectations, and a culture of busyness framed as intellectual seriousness. COVID and AI didn’t destroy these structures; they exposed their cracks. When educators talk about decline, students often hear blame. Yet many of the conditions Mintz describes, economic precarity, exhaustion, fragmented attention are not student inventions. They are structural realities our pedagogy has not kept pace with.

I don’t have the answers. And I think pretending we have them is part of the problem. But I do feel confident about involving students directly in this conversation. Student voices are talked about constantly, but students themselves are rarely talked with. I imagine that must be frustrating for them, watching instructors and institutions make sweeping decisions that shape their academic lives and engagement with material that is, supposedly, meant to strengthen their critical thinking skills.

This is a feeling I’ve been paying attention to for years. When I started my research five years ago on media literacy (specifically social networking sites) and wellbeing, I was motivated by the same question I’m bringing to my classroom now: What do students actually think? Not what do scholars assume. Not what do instructors predict. Not what does the data tell us in aggregate. But what do actual students, in actual classrooms, say about these issues?

That research ended up centering student voices across dozens of classrooms, and the findings revealed something important: students are more than willing to talk about topics like social media, wellbeing, and now AI. They have clear ideas about how these forces shape their learning, their mental health, and their identities. They bring nuance to conversations that dominant narratives often flatten. And they are, unsurprisingly, the ones living closest to the consequences of whatever policies, assumptions, or pedagogical shifts we adopt.

That work is set to publish with Palgrave Macmillan in 2026—one of the many reasons I’ve been so absent on Substack lately. Book deadlines, teaching, course redesigns, grant proposals, and new research projects all tend to collide at the same time. But returning to this space now feels grounding. It reminds me of why I write publicly in the first place: to think with people, not just at them.

Which brings me back to my students.

This post is rooted in a belief I hold firmly and return to often: student voices are essential to the AI conversation. Not optional. If we want to understand how AI is reshaping writing, learning, assessment, creativity, or motivation, we cannot keep centering only instructor anxieties or institutional policy debates. We have to listen to the people who live the reality of our classrooms every day.

I’m genuinely excited to see what students come up with next week—their critiques, their pushback, their insights, the angles I haven’t considered yet. And my next post will center their responses (with permission and anonymized, of course). I have a feeling they’re going to challenge all of us—including me—in the best possible way.

TTFN,

Dr. Sydney

Citations

Alam, A. (2023). Harnessing the power of AI to create intelligent tutoring systems for enhanced classroom experience and improved learning outcomes. In Intelligent communication technologies and virtual mobile networks (pp. 571-591). Singapore: Springer Nature Singapore.

Subscribe now

“AI is rapidly evolving. We don’t know what it will mean for learning yet. Let’s just be open, curious, and flexible.”

It’s the academic version of “We’ll figure it out later.”

And while I understand the impulse, especially cause of the fast-paced changes of AI. However, these kinds of policies aren’t neutral. They confuse students. They put instructors in impossible positions. And worst of all, they signal that the institution doesn’t actually know what learning looks like in the age of generative AI.

The problem isn’t uncertainty itself. The problem is pretending that uncertainty is a pedagogy.

When “Openness” Becomes Ambiguity

In the past year, I’ve seen more syllabi and institutional policies than I can count that fall back on some version of:

“AI tools like ChatGPT may be used when appropriate. Please discuss with your instructor what counts as appropriate use.”

I have even done this to some extent. Mostly because it sounds empowering. “Let’s talk about it together!” seems better than “you must do things my way.” But in practice, it still pushes all interpretive labor onto students and instructors. Students don’t yet know what responsible AI use looks like in academic writing. Faculty don’t agree on what counts as “appropriate.” And when everyone is told to “decide together,” no one actually knows what the rules are.

In my own courses, students have told me outright: “I want to follow the rules. I just don’t know what they are.”

This is what happens when “openness” becomes a euphemism for “we don’t want to commit.”

AI Literacy Isn’t Common Sense

Part of the reason this open-ended approach fails is that it assumes students already have the literacy to make informed decisions about AI use. But they don’t. And that’s totally okay. We shouldn’t expect them to.

Most students enter my classroom with either exaggerated confidence (“I know how to use ChatGPT to make an essay!”) or total uncertainty (“I’ve never touched it because I’m afraid I’ll get in trouble”). What they don’t have is a framework for why or when to use AI tools responsibly, or even the vocabulary to describe what these tools actually do.

When we invite them to co-write policies without first scaffolding their understanding, we aren’t democratizing the classroom. We’re setting them up to guess at values they haven’t yet been taught to name.

The Mirage of “Collaborative” Policy-Making

I’ve written before about co-writing AI policies with students, and I still believe in it. However, what I am starting to learn is that process only works if it’s not a free-for-all. It requires structure, critical literacy, and shared definitions.

Otherwise, what universities call “collaboration” is really surrender.

When I first piloted an AI policy co-writing framework, I made a critical mistake: I started by asking, “What do you think counts as ethical or unethical AI use?”
Students wrote their answers, but I could already tell the question I asked was a mistake. Not because they didn’t care, but because the question was too abstract.

So, next time we write policies, I am going to approach things differently. I will give them concrete writing tasks:

• Use AI to brainstorm topics. Assess and analyze this work.

• Use AI to rewrite an entire draft. Assess and analyze this work.

• Use AI to check grammar. Assess and analyze this work.

• Use AI to mimic a professor’s feedback. Assess and analyze this work.

Then ask: Which of these align with learning? Which replace it?

That changed everything. Students could engage critically because they had something tangible to reason about.

A “co-written” AI policy shouldn’t be a blank page. It should be a scaffolded conversation built on shared understanding, not shared confusion.

Clarity Isn’t Control

Some educators may avoid setting firm AI boundaries because they fear being authoritarian, as if saying “no” to certain uses of ChatGPT makes them anti-innovation. But clarity isn’t the opposite of curiosity. It’s what makes curiosity possible.

Think about any learning environment: rules are what allow for experimentation. A chemistry lab doesn’t say, “Let’s be open and flexible about which chemicals you mix.” A writing course shouldn’t say, “Let’s be open and flexible about what authorship means.”

When I define clear parameters—for instance, that AI can be used for brainstorming but not for final drafting—students don’t rebel. They relax. They understand where the boundaries are, and that gives them room to explore within them.

They also start asking better questions: “Can I use ChatGPT to identify bias in a source?” or “What’s the difference between paraphrasing and rewriting?”

These are learning questions, not compliance questions.

Concrete Doesn’t Mean Static

Of course, the tricky part is that AI changes weekly. So how can we be concrete about something so fluid?

Here’s how I’ve started to think about it:
Be flexible about tools, but concrete about values.

Students don’t need us to predict the future of AI. They need us to articulate the principles that guide our choices. That might be things like:

• Transparency: Always disclose when AI is used.

• Integrity: Use AI to assist thinking, not replace it.

• Learning: Choose methods that strengthen your own skills.

When students internalize these values, they can adapt them to whatever new tool emerges next semester: Claude, Gemini, Perplexity, or something we haven’t heard of yet.

A good AI policy, like a good syllabus, isn’t a list of prohibitions. It’s a shared framework for reasoning through change.

Students Need Language, Not Guardrails

When I ask students to describe how they used AI in an assignment, I’m struck by how little vocabulary they have to explain it. They say things like, “I used ChatGPT to make it better” or “I just asked it to fix it.”

That’s not dishonesty, it’s linguistic poverty.

Students need words like prompt engineering, bias, training data, hallucination, model output, authorship, voice, iteration. Without that vocabulary, they can’t meaningfully reflect on their own process.

One of the most effective things we can do right now isn’t to finalize policies, it’s to build shared language for talking about AI. Once students can describe how they’re using it, they can start to think critically about why.

A Closing Thought

AI in education is moving fast, but clarity and care don’t expire. When we stop hiding behind “openness” and start naming what we actually value—transparency, integrity, learning—we stop reacting to technology and start shaping how it’s used.

The future of teaching with AI doesn’t require omniscience. It requires courage to draw lines, explain why they matter, and revise them together when the world shifts again.

Subscribe now

Three days ago, a new revelation was made in relation to AI and hallucination. Published on Medium, the author argues that so-called hallucinations are better described as distance calculations gone wrong. Large language models don’t “dream” in the human sense; they’re calculating the most likely next word based on statistical distance in high-dimensional space. When there’s a gap in the data or ambiguity in the query, they don’t panic—they just guess. And sometimes that guess lands far from the truth.

That’s a much less mystical framing: not spooky visions, just math under strain. Perhaps the most important part of distance calculations gone wrong? They’re impossible to avoid, and they will always be part of the AI system. So, what does this mean for how we think about writing, especially research writing, in the age of AI.

What AI Errors Reveal About Research Writing

When you ask an AI to generate a research paper, you’re not tapping into a deep well of understanding. You’re prompting a pattern-matching engine. And in contexts that demand precision—like scholarship—that’s risky.

Here’s why:

1. Gaps in the training data
   If your paper requires an obscure citation, nuanced historical detail, or niche dataset, the AI may not have “seen” it before. So it synthesizes something plausible. That’s not creativity; that’s miscalculated distance.

2. Overconfidence as default
   Even when the model is uncertain, it outputs an answer with fluency and polish. Unlike a student who might say, “I’m not sure, I need to check,” the AI rarely withholds.

3. Mismatch with scholarly standards
   Academia thrives on careful sourcing, reproducibility, and intellectual honesty. AI thrives on sounding right. Those are not the same thing.

4. Fluency masks failure
   The biggest trap: a sentence that sounds good but is conceptually empty or factually wrong. Unless the reader has a strong grasp of the subject, they may not notice.

In other words, AI isn’t misbehaving when it invents a citation. It’s doing what it was built to do—fill gaps with probabilities. The problem is that research writing doesn’t reward “close enough.”

From Research Papers to “Workslop”

This problem doesn’t end in the classroom. It’s already spilling into the workplace.

Harvard Business Review recently published a piece with a striking headline: “AI-Generated Workslop Is Destroying Productivity.” The term “workslop” captures the deluge of low-quality drafts, memos, and reports that AI can churn out. This content looks complete but requires significant cleanup.

The promise of AI in the workplace was speed. But speed without accuracy or judgment can become a time sink. Companies are finding that workers spend hours editing, verifying, and restructuring AI outputs. In fact, “a recent report from the MIT Media Lab found that 95% of organizations see no measurable return on their investment in these technologies.” The result: productivity gains vanish, and sometimes efficiency actually drops.

Why? Because workers who lack strong foundational skills—writing, critical thinking, domain knowledge—struggle to tell good output from bad. They accept mediocrity, or worse, fail to recognize errors at all. AI becomes a crutch for rather than a tool.

The Education Connection

Education can step in here, not as an abstract virtue, but as the bedrock of effective AI use.

• Domain knowledge lets you see when the AI veers off course.

• Critical reading teaches you to interrogate claims, not just nod along.

• Writing skills help you shape, refine, and structure ideas instead of settling for the AI’s first draft.

• Statistical literacy allows you to understand uncertainty and bias in model outputs.

• Ethical reasoning reminds us that not everything that can be automated should be.

Without these foundations, AI is less like an assistant and more like a saboteur: flooding workflows with polished nonsense. With them, AI can become a powerful partner—drafting outlines, brainstorming counterarguments, accelerating revision cycles, or simulating alternative perspectives.

Beyond Tools: A Call for Integration

So where does this leave us? A few takeaways for both students and professionals:

1. Rethink metaphors. AI doesn’t “hallucinate.” It calculates and misses. Treating it as mystical makes us too trusting; treating it as math keeps us critical.

2. Use AI as scaffolding, not oracle. Ask it to help you brainstorm or reframe, but keep final judgment—and truth claims—firmly human.

3. Invest in foundational education. The better we are at reading, writing, and reasoning, the better we can harness AI without drowning in “workslop.”

4. Redefine productivity. It’s not about how many words we generate—it’s about how many useful words survive revision.

Building a Future With, Not Against, AI

The conversation around AI too often splits between hype and fear: AI will either save us or destroy us. The reality is more mundane but more urgent: AI will amplify whatever skills we bring to it.

If we bring rigor, discernment, and foundational knowledge, it can accelerate high-quality work. If we bring complacency and overreliance, it will bury us in polished junk.

Education isn’t just preparation for a world with AI. It’s the difference between being empowered by these tools and being undermined by them.

So,

the next time someone says AI “hallucinated,” remember: the real hallucination would be thinking we can outsource thinking itself.

My students are finally starting their research papers. For many of them, this will be their first college research paper, and I already know AI needs to be discussed as we move forward. I’ve seen platforms like ChatGPT write research before—I’ve even seen students who clearly used it. And yet, when we get to conferencing together to discuss their “perfect” product, it is clear the student feels entirely dispossessed of their own work. Even when AI writes the first draft, my students stall, unable to write the rest, because the intellectual groundwork wasn’t theirs.

This student’s moment of crisis is our shared dilemma. In a faculty meeting a few weeks back, a colleague voiced how we aren’t merely facing a new cheating tool; we are grappling with a profound existential challenge to the undergraduate learning model. If an LLM can simulate the product of academic labor in seconds, how do we justify the four-year process we demand?

I couldn’t agree more. The problem isn’t the technology; it’s our traditional assessment model. If a core requirement of a degree—the first research paper—can be outsourced to AI, the fault lies with the assignment, not the student. Our task isn’t to police the technology, but to redefine where the enduring value of a college degree truly lies.

The Obsolescence of Specific AI Skills

It’s tempting to pivot our curricula to teaching students proprietary AI skills: the perfect “prompt engineering” or the most current platform interface. We might spend a week showing them how to use specific commands.

This is a pedagogical dead end.

We can’t predict where Generative AI technology will be in four years when our current students graduate in 2029. The specific tools we teach today will likely be obsolete, replaced by faster, more integrated, and potentially invisible systems. Teaching students to use this specific AI is a futile exercise in chasing a constantly moving, temporary target.

Our focus, then, must pivot from teaching tool-specific competency to nurturing durable critical thinking—the cognitive scaffolding that allows a student to adapt to any future tool. The value of a degree is no longer in the information students accumulate, but in the intellectual friction they learn to tolerate, leverage, and master.

The Value of Friction: Reclaiming the First Research Project

The first research paper is supposed to be a crucible. It’s a moment where a student learns the discipline of intellectual ownership. To preserve this, we must build a First-Draft Firewall, identifying the specific, high-friction stages of research that are actively damaged by AI use and making them mandatory human-powered processes.

Here are the three critical thinking moments that must remain human-centric, and the assignment shifts you might consider:

1. Defining the Question (The Why)

The single most important, non-delegable skill in research is defining the problem.

A novice researcher using AI will inevitably ask: “Write me a paper on the effects of climate change.” The AI response will be a generic, five-point summary. The student avoids the essential, painful labor of narrowing, contextualizing, and defining a genuine question.

The Durable Skill: Intellectual Autonomy and Scoping.

The Assignment Shift: We need to grade the question before the paper is even written. Instead of banning AI, let’s use it to highlight the stakes.

• Pre-Assignment: Instruct students to generate five different “research paper titles” or “research angles” on their chosen topic using ChatGPT.

• The Graded Task: Students must then write a 500-word reflective memo that critiques these AI-generated options. They must explain, with justification, why four of the titles are too broad, too generic, or ethically inert, and why their final, human-crafted question is the most original, manageable, and specific for the scope of the class.

This shifts the labor from content generation to critical judgment. A skill AI cannot replicate.

2. Evaluating a Claim’s Source and Context (The Where)

AI is a statistical machine trained on data; it is not a scholar trained in epistemology (the study of knowledge). It synthesizes plausible-sounding text, but it cannot perform the intellectual work of distinguishing a foundational methodology paper from a retracted study, or a primary source from a tertiary summary.

The Durable Skill: Skepticism and Source Hierarchy.

The Assignment Shift: Integrate source validation and contextualization into the early stages of the project.

• The Moment to Watch For: The instant a student asks AI for a citation or a literature review. This removes the necessity of scrolling through databases, reading abstracts to distinguish a methodology, and building the disciplinary mental map.

• The Graded Task: Implement a “Source Scrutiny Log.” For their first five required sources, students must provide not just the citation and a summary, but also a Critique of Provenance: Why does this source have authority? What is the author’s institutional affiliation? How is this source connected to (or arguing against) two other key texts in the field?

This forces students to perform the slow, contextual work of research, making the AI-generated “instant citation list” useless.

3. Analyzing the Data-to-Conclusion Link (The How)

The greatest risk of AI for the novice researcher is that it replaces the analytical journey with a destination. Students bypass the messy, difficult work of moving from evidence to insight.

The Durable Skill: Argumentation and Analytical Scrutiny.

The Assignment Shift: Use AI as an adversary, not a collaborator. The critical work must be in the revision and defense of the original human thought.

• The Graded Task: Require a Defensive Drafting Stage. Students must first write their conclusion (based on their own analysis of their sources) without AI. They then run their draft conclusion and evidence through an LLM, asking it to act as a critical peer reviewer and generate two counter-arguments. Students then submit their final conclusion alongside a 300-word Defense Memo explaining how they addressed or refuted the AI’s critique.

This transforms the AI from a substitute for deep work into a challenging sparring partner, thereby strengthening the student’s own human capacity for reasoning.

Earning the License to Think

We are not teaching specific prompt skills; we are teaching students how to be the human editor, the chief skeptic, and the intellectual owner of their work.

In four years, the technical job functions that rely on repetitive logic will be augmented or automated. What will be left is the human capacity to define, question, judge, and create the original idea. By designing assignments that prioritize the cognitive process over the polished product, we ensure that our students earn their degree by mastering the valuable struggle. This is the license to think that no machine can confer.

But I feel like this is getting to be a tired song. We get it. AI is “revealing” to us that our system is flawed. But honestly, many scholars knew this to be true and were arguing about it for years before AI—Paulo Freire’s work on critical pedagogy, bell hooks’ insistence on engaged teaching, John Dewey’s focus on democratic classrooms, and more contemporary critiques of neoliberal education from scholars like Henry Giroux and Cathy Davidson all made it clear that education had drifted away from its core values long before ChatGPT arrived.

Subscribe now

So instead of focusing on the past, or even looking too far forward, I want to be present with AI. Where are we at now?

What the Research Shows

Recent research gives us a clearer snapshot of this moment. Ng, Chan, and Lo (2025), writing in Computers and Education: Artificial Intelligence, highlight how teachers are cautiously optimistic about generative AI but remain concerned about readiness: schools lack clear guidelines, teachers lack confidence in their AI competencies, and students lack literacy and ethical grounding. Their study suggests that while teachers see the potential for generative AI to personalize learning, reduce workloads, and foster engagement, the infrastructure and support systems lag far behind. Professional development, policy frameworks, and student AI literacy are simply not keeping pace with the speed of technological change.

Similarly, Eden, Onyebuchi, and Adeniyi (2024) stress that AI in education is a double-edged sword. On one hand, it can enhance personalization, automate administrative tasks, and expand access through immersive tools like AR and VR. On the other, it raises red flags about equity, privacy, and bias. They argue that the ethical frameworks we adopt now will shape whether AI reinforces inequities or helps close them. Without thoughtful oversight, the very same algorithms that personalize learning could perpetuate the same structural disadvantages educators have been struggling against for decades.

Taken together, these perspectives suggest that “being present” with AI requires us to resist simplistic narratives. It’s not enough to say “AI will revolutionize everything” or “AI will destroy education.” The reality is more complicated: AI is already here, embedded in classrooms, grading tools, writing assistance apps, and administrative systems. The task before us is less about deciding whether AI belongs in education and more about how to guide its use in ways that align with our deepest educational commitments.

An Educational Approach: AI as a Pedagogical Mirror

One way to frame this moment is to see AI not just as a tool but as a mirror. When Ng et al. (2025) report that teachers lack confidence in their AI competencies, what they are really describing is a deeper tension: our systems of teacher preparation and professional development were not designed to keep pace with rapidly evolving technologies. When Eden et al. (2024) warn about data privacy, what they highlight is how much we’ve allowed educational technologies to outpace ethical safeguards.

Rather than treating AI as an external disruption, we might view it as a spotlight on the unfinished work of education. It forces us to revisit foundational questions:

• How do we teach students to be critical consumers and creators of information?

• How do we balance efficiency with equity?

• How do we prepare teachers not only to use new tools but to question them, adapt them, and set limits around them?

This reframing matters because it keeps us grounded in pedagogy rather than technology. AI doesn’t erase the need for critical pedagogy, student-centered learning, or democratic classrooms. In fact, it intensifies those needs. If generative AI can draft essays in seconds, then the value of writing instruction shifts toward process, reflection, voice, and critical evaluation. If AI can provide adaptive quizzes, then the educator’s role becomes less about delivering practice problems and more about helping students interpret feedback, connect ideas, and build metacognitive awareness.

In other words, AI changes what we must emphasize in teaching, but it does not replace the human relationships and reflective practices at the heart of learning.

Holding the Tension

Being present with AI means holding the tension between opportunity and risk. Teachers see exciting possibilities: personalized feedback, streamlined grading, lesson planning assistance, even creative collaboration with students. At the same time, they worry (rightly) about privacy breaches, bias, plagiarism, overreliance, and the erosion of critical thinking.

This tension doesn’t mean we are stuck. It means we are in a formative moment, where the habits, norms, and policies we establish will shape the role of AI in education for years to come. If schools adopt AI reactively, allowing corporate tools to set the terms, then equity and ethics will be afterthoughts. But if educators, administrators, and students actively shape how AI is integrated, then it can become a catalyst for reclaiming education’s core values.

Practical Applications: Tools for Being Present with AI

So what does this look like in practice? How do we as educators “be present” with AI rather than stuck in either nostalgia or speculation? Here are a few concrete approaches:

1. AI Literacy Modules for Students

   • Teach students how AI tools like ChatGPT actually work (e.g., large language models, bias in training data, limitations of outputs).

   • Have them practice evaluating AI-generated content for accuracy, bias, and credibility rather than simply using it.

   • Pair AI use with reflective prompts: What did the AI get right? What did it miss? How did it shape your own thinking?

2. Process-Based Writing Assignments

   • Instead of banning AI in writing, redesign assignments around process. Require students to submit brainstorming notes, drafts, revision memos, and reflections on where and how they used AI.

   • This keeps the focus on metacognition and self-awareness, skills AI can’t automate.

3. Teacher Professional Development on Prompting and Ethics

   • Many teachers feel unprepared to use AI effectively (Ng et al., 2025). Schools can host workshops on “prompt engineering for the classroom” alongside discussions of ethical concerns (privacy, plagiarism, equity).

   • The goal is not just technical fluency but critical confidence: knowing when and how to use AI and when to set it aside.

4. Classroom Dialogues on Equity and Bias

   • Use examples of biased AI outputs as teaching moments. Invite students to question why bias appears, how it relates to data collection, and what ethical responsibilities we have when using such tools.

   • Connect these conversations to broader lessons about media literacy, social justice, and digital citizenship.

5. Small-Scale Pilot Projects

   • Rather than overhauling entire courses, educators might try limited pilots: using AI to generate quiz questions, co-create rubrics, or provide first-draft feedback.

   • Collect student and teacher reflections on what worked and what didn’t, and use this evidence to shape future policies.

Conclusion

To be present with AI is to neither romanticize the past nor rush blindly into the future. It is to sit with the complexity of this moment—acknowledging both the excitement and the unease. The studies by Ng et al. (2025) and Eden et al. (2024) remind us that the real work is not about whether AI belongs in education. It already does. The work is about how we guide its use—ethically, equitably, and intentionally—so that it supports rather than undermines the core values of learning.

If we use this moment well, AI can become more than a technological tool. It can be a catalyst for re-centering education around critical thinking, reflection, equity, and care. But that requires us, as educators, to engage with AI not just as consumers but as shapers, critics, and co-designers.

The question isn’t whether AI will change education. It already has. The real question is whether we will let those changes drift us further from our values—or whether we will harness them to bring us closer to the kind of education we have always hoped for.

References

Davidson, C. N. (2017). The new education: How to revolutionize the university to prepare students for a world in flux. Basic Books.

Dewey, J. (1916). Democracy and education. Macmillan.

Eden, C. A., Onyebuchi, N. C., & Adeniyi, I. S. (2024). Integrating AI in education: Opportunities, challenges, and ethical considerations. Magna Scientia Advanced Research and Reviews, 10(2), 6–13. https://doi.org/10.30574/msarr.2024.10.2.0039

Freire, P. (1970). Pedagogy of the oppressed. Continuum.

Giroux, H. A. (2014). Neoliberalism’s war on higher education. Haymarket Books.

hooks, b. (1994). Teaching to transgress: Education as the practice of freedom. Routledge.

Ng, D. T. K., Chan, E. K. C., & Lo, C. K. (2025). Opportunities, challenges and school strategies for integrating generative AI in education. Computers and Education: Artificial Intelligence, 8, 100373. https://doi.org/10.1016/j.caeai.2025.100373

Subscribe now

Last week’s discussion of grades and AI sparked big questions about the purpose of higher education. This week, I want to push that conversation further by asking: what if AI could be even smarter than a professor?

This is by no means an idle question. In the past two years, AI has moved from the margins of higher education into the center of debates about teaching, assessment, and even the purpose of the university itself. Professors are no longer the unquestioned gatekeepers of knowledge. Students can consult ChatGPT in the middle of the night and get feedback faster than we could ever hope to return an email. They can ask an AI tutor to explain Aristotle or derivatives or APA formatting and receive an instant response. The uncomfortable truth is that in many narrow ways, AI already is “smarter” than us: quicker, more responsive, and (at least on the surface) endlessly patient.

When Professors Feel Replaceable

N. Weimann-Sandig (2024) describes how faculty in Germany are grappling with precisely this disruption. She notes that many professors feel their authority “sensitively disturbed” when AI can handle technical questions or provide feedback on scientific writing. If knowledge transmission is no longer our monopoly, what is left of our role?

Her answer is to reframe teaching as co-constructive: not professors depositing knowledge into students (a very old model) but teachers and learners constructing meaning together. In this model, AI is not a competitor but a tool that can support exploration. Professors remain essential not because we know more facts, but because we help students interpret, contextualize, and use those facts meaningfully.

That’s a beautiful vision. And yet, beneath it lies a fear many of us carry quietly: if AI can do my tasks, how do I prove my worth in this job?

The Five Levels of an “AI Professor”

Cheolkyu Shin and colleagues (2024) take this fear head-on by mapping out a framework for the “AI professor.” Inspired by the automation levels of self-driving cars, they outline five stages—beginning with AI as a simple assistant (today’s reality) and ending with a fully automated professor at Level 5. At that final stage, human faculty wouldn’t even need to supervise; AI would design, teach, assess, and advise entirely on its own.

This is both fascinating and terrifying. The authors admit that technical barriers make Level 5 still hypothetical. But they also stress that the real danger isn’t whether AI can do all the tasks of a professor—it’s whether universities, under pressure to cut costs and boost efficiency, will be tempted to try.

Here’s the catch: Shin et al. emphasize that professors do far more than deliver content. They list five competencies: expertise, class structuring, teaching, assessment, and communication. AI may be able to automate aspects of each, but the relational, ethical, and human elements remain stubbornly beyond its reach. What’s at stake isn’t whether AI can be programmed to grade essays or answer calculus questions. It’s whether we, as institutions and as a society, believe that education is only about efficient delivery or instead about mentorship, formation, and the messy process of becoming.

“Better Than My Professor?”

Stefano Triberti (2024) frames the tension with a blunt question: what if AI is better than my professor? His analysis of conversational pedagogical agents—chatbots that can answer questions, give feedback, and even draw on motivational theories to encourage students—is sobering. Students often prefer the speed and responsiveness of AI. In distance education especially, a chatbot that answers at midnight can feel more supportive than a professor juggling 200 emails.

But Triberti argues that the real promise of AI is not in replacing professors but in relieving them. If chatbots can handle administrative questions (“When is the deadline?” “Where do I upload my file?”) and basic clarifications, professors can spend more time on higher-level engagement: deep discussions, mentoring, and guiding students through ethical and interdisciplinary questions. AI, in other words, might be better at certain things, but those aren’t the things that define the heart of teaching.

Share

The Worth Question

Still, the emotional undercurrent remains: how do we prove our worth in this new environment? For professors, worth has always been precarious. We measure ourselves in student evaluations, publication counts, grant dollars, and the constant hustle to justify our positions. The intrusion of AI raises the stakes: now we are not only competing with each other, but also with tools that never sleep, never complain, and can scale infinitely.

I’ve felt this tension in my own teaching. When a student asks, “Why can’t I just ask ChatGPT?” the subtext is, “Why do I need you?” My instinct is to respond with all the ways I’m different: I can understand nuance, challenge assumptions, provide mentorship. But the truth is, proving my worth isn’t about drawing a defensive line between “what AI does” and “what I do.” It’s about leaning into what only a human relationship can provide.

And yet, that’s hard. It’s one thing to know intellectually that mentorship matters; it’s another to believe it counts as much as a published paper or a filled classroom when tenure committees or administrators are making decisions. The worth question is not just about identity—it’s about economics.

Shifting the Metrics

What if we changed the way we measure professors’ value? Instead of evaluating us solely on output—syllabi produced, classes taught, papers published—what if we rewarded relationships built? What if the most important Key Performance Indicator wasn’t student pass rates, but whether students felt known, challenged, and supported in ways AI never could?

The irony is that much of the research I’ve cited already points us there. Weimann-Sandig’s co-constructive model depends on authentic human relationships. Shin et al. admit that even Level 5 AI professors will still require human oversight for ethical and relational reasons. Triberti concludes that professors are indispensable for mentoring and guiding students beyond technical knowledge.

So maybe the challenge isn’t that AI will be “smarter” than us. It’s that universities may not yet have the courage to value the parts of our job that machines can’t do.

A Future of Collaboration, Not Competition

If AI continues advancing (and it will) the question isn’t whether it will outsmart us in certain tasks. It already has. The deeper question is whether we as educators will cling to proving our worth by being faster, more knowledgeable, more efficient—or whether we’ll reimagine our worth around what only humans can offer: empathy, ethical discernment, contextual wisdom, and the courage to accompany students through uncertainty.

In that sense, the arrival of “AI professors” might be less of a threat than an invitation. An invitation to let go of the illusion that our worth lies in grading stacks of essays at 2 a.m., or answering every factual question with perfect recall. And an invitation to double down on what drew many of us to this vocation in the first place: being part of the human project of learning together.

So yes, AI might someday be “smarter” than a professor. But teaching has never just been about information; it’s about transformation.

TTFN,

Dr. Sydney

But I think we’re asking the wrong question. Instead of asking how to stop them, we should be asking: Why are they turning to AI in the first place?

And here’s the uncomfortable truth: the answer has less to do with laziness and a lot more to do with grades.

The Problem Isn’t AI, It’s Grades

Back in 2009—long before ChatGPT—researchers published a study in the Journal of Business Ethics that looked at why students cheat. They surveyed both business and non-business majors, and over 60% admitted to cheating at some point. But here’s the kicker. The top reason wasn’t hatred of the subject, or even apathy. It was the “desire to get ahead.”

In other words, students weren’t cheating because they didn’t care. They were cheating because they cared too much about performing in a system where points, grades, and credentials often outweigh actual learning. Fast forward to 2025, and AI is just the newest tool in that same old game. The pressure cooker hasn’t changed. The shortcuts just got more powerful.

Which brings me to what I think is the real solution: stop grading in ways that make students feel like their worth depends on winning the system.

Contract Grading as the “AI Antidote”

This is where Asao Inoue’s work on labor-based grading contracts changed my teaching life.

The idea is simple but radical: instead of grading students on whether their work meets a polished, professor-defined standard, you grade them on their labor—the effort, process, and reflection they bring to the course. Show up, engage meaningfully, revise, take risks, reflect, and you succeed.

At first, this felt terrifying. Aren’t students going to slack off? Won’t classrooms devolve into chaos?

The opposite happened.

When students no longer felt the need to write for my approval—or for a letter grade that could make or break their GPA—they wrote for themselves. They took risks. They admitted when they were confused. They experimented with AI tools honestly, instead of sneaking them under the table.

Contract grading didn’t just change how they wrote. It changed why they wrote. Essentially, when grades stop being the carrot or the stick, AI stops being a cheat code.

The First-Week Policy Experiment

This summer, I decided to go one step further. Instead of writing an AI policy for my class, I asked my students to write it with me.

On the very first week, I posted a prompt on our course site:

“Imagine this isn’t a classroom. It’s a writing group. We’re here to grow as thinkers, writers, and community members. One of the tools we might use is AI—but what should that look like? What should the boundaries be?”

Students responded on a shared Padlet board. Some wrote bullet points, others paragraphs, some just posted questions. Their comments blew me away:

• “AI can be a tool, not a crutch. If it does the hard part for you, you won’t grow.”

• “It’s like Grammarly but on steroids. Okay if you’re still learning from it.”

• “What’s the point of taking a writing class if you’re not going to try writing?”

• “AI is good for outlines and ideas, but not for actual voice.”

• “We need to be transparent. If you used AI a lot, say so.”

Across all the responses, a shared philosophy emerged: AI should assist learning, not replace it.

I compiled their ideas into a living class policy. The gist looked like this:

Ethical Uses of AI

• Brainstorming ideas

• Clarifying confusing concepts

• Grammar/sentence-level revision

• Organizing ideas or outlines

• Asking for growth-oriented feedback

Inappropriate Uses of AI

• Writing full essays or assignments

• Answering quizzes or graded tasks

• Replacing your original voice

• Using heavily without disclosure or critical thought

Transparency
“If you use AI in a significant way, include a quick note in your author’s memo. Example: ‘I used ChatGPT to help brainstorm my topic and reword the intro paragraph.’”

That was it. Nothing fancy. But the effect was huge: students felt seen, they felt trusted, and they felt like the rules were something they owned, not something imposed on them.

Why This Works Better Than Policing

Let’s be honest: if a student really wants to use AI to write an essay behind my back, they’ll find a way. No amount of surveillance software will change that.

But surveillance does breed mistrust. It drives AI use underground. It tells students: I expect you to cheat, so I’m watching you.

Contrast that with contract grading plus co-written policy:

• Trust instead of suspicion. Students know their voices matter.

• Transparency instead of secrecy. Students disclose AI use without fear.

• Ethical reasoning instead of rule memorization. Students learn to evaluate what’s helpful vs. harmful.

This isn’t just about AI. It’s about the kind of citizens we’re cultivating. Do we want students to graduate knowing how to dodge detection? Or knowing how to think critically and ethically about the tools they’ll inevitably use in life and work?

A Culture Shift

A few weeks ago, I came across a line from educator Maurie Beasley that stuck with me:

“AI didn’t kill deep reading. We did. With packets instead of exploration. Five-paragraph essays instead of ideas. ‘Did you cite the author’s middle name?’ instead of ‘What did this story mean to you?’”

That’s exactly it.

The rise of AI isn’t a crisis of technology. It’s a crisis of pedagogy. If we’ve trained students to perform for points, we shouldn’t be surprised when they reach for shortcuts.

But when we design for curiosity, trust, and growth, the need for shortcuts evaporates. Students stop hiding. They start engaging. And they begin to see AI not as a way to cheat the system, but as one tool in a much bigger toolbox of learning.

Try This in Your Own Classroom

If you want to try co-writing an AI policy with your students, I’ve put together a free plug-and-play worksheet you can adapt for your own class.

👉 Download the worksheet here

All I ask is that you share it with another educator who might benefit from the practice. Imagine if, instead of a patchwork of AI “bans” and surveillance policies, we had classrooms everywhere where students were trusted to co-create their own rules.

Final Thoughts

Not grading won’t magically erase every AI problem. But it changes the conditions that make misuse tempting. When grades stop being the focus, learning has a chance to be.

And when students co-create the rules, they don’t just follow them—they own them.

If AI is the spark that finally pushes us to rethink grading, maybe that’s not a crisis at all. Maybe it’s an opportunity to build classrooms that are less about getting ahead and more about growing together.

Over to you: Are you experimenting with contract grading, ungrading, or co-created policies this fall? What’s working—or not working—for you? Drop your thoughts in the comments.

TTFN

-Dr. Sydney

And that’s how the current state of AI writing feels.

Every day, dozens of new posts flood our feeds: articles from think tanks, press releases from companies, hot takes on LinkedIn, research papers, Twitter threads, Substack debates. Each one promises to tell us the thing we need to know. The result? We’re all standing in the middle of a digital library, trying to hold back the tidal wave.

The reality is—just as in a real library—you can’t read it all. You shouldn’t even try. The trick is learning to filter. Here are some of the filters I’ve found most helpful:

Subscribe now

1. Filter for clarity, not hype

If the headline leans on extremes—“AI will save education forever!” or “AI will destroy all human creativity!”—skip it. The truth is almost always more nuanced. Instead, prioritize writing that slows down, explains how something works, and gives tangible examples.

2. Check the source’s stake in the game

Ask yourself: who benefits from me believing this? If a company is selling a tool, I read with a grain of salt. If it’s a classroom teacher testing a new assignment, I pay closer attention. Both can be valuable, but it helps to know the writer’s angle before you invest time.

3. Prioritize synthesis over novelty

This is where the library metaphor stings most. When I walk into a library and feel the ache of not being able to read everything, I remind myself that what I really need isn’t more books but books that bring ideas together.

The same goes for AI. Not every “new” update is worth chasing. What matters are the pieces that connect the dots—the writers who step back, compare findings, highlight patterns, and make sense of the bigger picture. That’s where learning compounds. That’s where overwhelm softens into insight.

4. Follow a few trusted curators

You don’t need hundreds of voices. Find a handful of people who consistently do the work of synthesis, critique, and thoughtful experimentation. For me, that list includes (whose Substack is a masterclass in cutting through noise), along with researchers and educators like . They each approach AI and education differently, but what unites them is their ability to filter first—then share what matters most.

5. Give yourself permission to miss things

This one is the hardest, and maybe the most important. In a library, you eventually accept that you’ll never read everything—and that doesn’t diminish the experience. With AI, too, it’s okay to let things go by. If it’s truly important, it will resurface. The internet has a way of recycling the ideas worth holding onto. If you’re really wanting to integrate AI into your pedagogy, I would recommend reading the way you would any new theory or idea (3-5 articles a week, mixing it up between peer-reviewed and opinion-based).

The bottom line: You don’t need to read everything to stay informed. You need to read the right things—and that requires a filter system that protects your time, energy, and attention. Think of it less like drowning in information, and more like choosing a few good books to carry out of the library.

In the comments, many instructors cheer. Finally! A way to outsmart the bots. While many other users throw sarcastic jabs, the general sentiment being “oh no, now you have to do your own work, boohoo.” But lurking just beneath is a more complicated question: yes, you can stop students from using AI this way. But should you?

I think the answer is no—and not because I believe students should outsource their work to ChatGPT, but because doubling down on surveillance undermines the very kind of learning we say we value.

1. The Problem with Automatic Distrust

At the heart of the screen-recording idea is an assumption: students cannot be trusted. If we accept that premise, then every policy we create will be built around catching, punishing, or blocking dishonesty. That might feel like a practical response to AI, but it comes at a real cost. It shifts the classroom relationship from collaboration to control.

We’ve seen this before. During the pandemic, remote proctoring software like ProctorU and Respondus Monitor became widespread, requiring students to submit to facial recognition scans, 360° room sweeps, and eye-movement detection. Critics argued that such surveillance disproportionately harmed students with disabilities, neurodivergent students, and students without private spaces. The unspoken message was clear: your word isn’t good enough. You must be watched to be believed.

When I hear instructors talk about “banning AI,” I often think about the missed opportunity. Instead of building environments where students can engage in honest, critical conversations about how these tools shape our work, we reduce everything to a cat-and-mouse game. Teachers invent restrictions, students invent workarounds. Nobody learns much except how to evade the system.

Subscribe now

I’ve written before about the power of co-creating AI guidelines with students. When students help define what counts as ethical use, they seem not only to understand the boundaries better, but they also buy into them. They move from fearing punishment to practicing judgment. And isn’t that what education is supposed to cultivate—the ability to think critically, ethically, and independently in contexts that don’t have clear rules?

Automatic distrust, by contrast, cuts those conversations off entirely. It says: “We don’t need your input, because we already assume you’ll cheat.” It closes the door on nuance, and with it, the possibility of genuine engagement with the ethical complexities of AI.

2. If They Wanted To, They Would

Let’s imagine, for a moment, that you do implement the screen-recording strategy. What happens next?

The comments on that Instagram post already give us a preview. Students suggest easy workarounds:

• Use ChatGPT on a second device, then retype everything on the “official” computer.

• Have a friend draft ideas, then paraphrase them during the screen-recorded session.

• Start with AI-generated outlines, then expand them into full paragraphs while recording.

In other words: if a student is determined to use AI, they will.

This mirrors decades of research on academic dishonesty, which consistently shows that surveillance-based deterrents have limited effects. McCabe, Butterfield, & Treviño found that students cheat less when they believe their instructors value integrity and trust them, not when instructors simply increase detection measures. Similarly, Lang synthesizes the research on cheating and reminds readers that cheating is increased by having a classroom with “an emphasis on performance,” “high stakes,” “an intrinsic motivation for success,” and “low expectations of success.” All of which are pertinent issues to address in the classroom and likely to be impacted negatively be simply banning or taking a top-down approach with AI.

Strict anti-AI hacks might technically prevent some instances of cheating, but they do little to address the underlying issue: why a student feels compelled to use AI dishonestly in the first place. Meanwhile, students who weren’t going to cheat are now burdened with a new layer of surveillance—recording themselves, managing files, uploading evidence—that treats them like potential criminals.

The result? A classroom climate of suspicion that helps no one.

3. The Missed Opportunity: Designing for Learning, Not Policing

This brings me to the third point, which I think is even more important than the first two: every hour we spend devising new ways to “catch” students is an hour we’re not spending redesigning our assignments to foster authentic learning.

The real question isn’t “How can we stop students from using AI?” It’s “How can we create learning experiences where outsourcing the work isn’t appealing in the first place?”

That might mean:

• Process over product. Build assignments around drafts, reflections, and peer reviews, so that the journey and process of writing matters as much as the final essay. Many writing scholars have already documented the importance of a process approach in the writing process.

• Transparency over secrecy. Ask students to document where they used AI (if they did) and to reflect on how it shaped their process. This doesn’t punish use—it makes it part of the critical inquiry.

• Choice over compliance. Give students flexibility in topics or formats so they feel invested in the outcome. When work feels personally meaningful, shortcuts are less tempting.

These shifts align with broader moves in writing pedagogy that emphasize critical digital literacy—not just teaching students to produce “authentic” text, but helping them engage with the technologies that mediate knowledge production.

When we lean into these design choices, AI becomes less of a threat and more of a teaching tool. Students begin to see that the value of their education isn’t in producing “AI-proof” paragraphs but in learning how to navigate technologies that will only become more pervasive in their professional and personal lives.

So, Should You Stop Students from Using AI?

You can. With enough surveillance and restrictions, you can force students to perform a narrow version of “authentic” writing. But that kind of learning environment comes at a cost: distrust, resentment, and endless arms races between instructor controls and student workarounds.

Or you can take a different approach. You can invite students into the conversation. You can design assignments that foreground process, reflection, and choice. You can treat AI not as an enemy to outwit, but as a cultural force worth examining together.

The first path builds compliance. The second builds trust.

And trust, in the long run, is the only real safeguard against cheating, because it reframes education not as something to “get around,” but as something worth doing in the first place.

Until next week. TTFN.

— Dr. Sydney

More than anything, it feels like AI has become the central focus of education. So, I wanted to know—at this particular point in society’s AI journey—what scholarship is getting the most attention. Here are the top five most cited articles I found on Google Scholar, with their citation counts included, and some ideas or responses I have on how these research insights work in the classroom.

Subscribe now

Abbas, M., Jam, F. A., & Khan, T. I. (2024). Is it harmful or helpful? Examining the causes and consequences of generative AI usage among university students. International journal of educational technology in higher education, 21(1), 10.

Cited by 346

In their 2024 article, "Is it harmful or helpful? Examining the causes and consequences of generative AI usage among university students," Abbas et al. present one of the first large-scale empirical investigations into how and why students use ChatGPT in academic contexts—and what consequences that use brings. Drawing on data from two studies (N=165 and N=494), the researchers developed a validated 8-item scale to measure ChatGPT usage and examined the psychological and academic factors associated with it. They found that students experiencing high academic workload and time pressure were significantly more likely to rely on ChatGPT, whereas students who were highly sensitive to rewards (i.e., grades) were less likely to use it—possibly due to fears of plagiarism penalties. Surprisingly, sensitivity to quality did not significantly predict usage, suggesting that students’ motives may be more about managing stress than improving their work.

The study also revealed that heavy use of ChatGPT was associated with increased procrastination, memory loss, and lower academic performance. In essence, while the tool may offer short-term relief from academic pressures, it may inadvertently foster habits that undermine cognitive engagement and long-term success. Importantly, ChatGPT usage mediated the relationship between stressors like workload and time pressure and these negative outcomes. Abbas et al. conclude with a call for educators and institutions to carefully consider how AI tools are integrated into learning environments—encouraging balanced use and redesigning assessments to promote critical thinking and independent work. This article has quickly become foundational in the academic conversation about AI in higher education, offering timely insights into both the promise and the perils of generative AI.

Last week, I wrote a piece titled “AI Proofing Your Classroom (Sort Of).” The premise of which was very rooted in what the above scholarship strives for: understanding students’ motivations for AI Usage and adjusting from there. What I found particularly interesting was Abbas et al.’s discovery that motivations for AI usage were likely not due to students wanting to improve their writing. It could be, in fact, that ChatGPT was never about “better” writing, and always about stress management. It’ll be interesting to see how the research in this area plays out in the future.

Stöhr, C., Ou, A. W., & Malmström, H. (2024). Perceptions and usage of AI chatbots among students in higher education across genders, academic levels and fields of study. Computers and Education: Artificial Intelligence, 7, 100259.

Cited by 157

In "Perceptions and Usage of AI Chatbots Among Students in Higher Education Across Genders, Academic Levels and Fields of Study" (2024), Stöhr, Ou, and Malmström offer one of the largest empirical snapshots to date on how students engage with ChatGPT and other AI chatbots. Drawing on survey data from nearly 6,000 (n = 5894) students across Swedish universities, the study reveals that ChatGPT was by far the most recognized and frequently used chatbot—regularly used by over 35% of students—while other tools like Bing AI and Bard lag behind in familiarity and adoption. Over half of student respondents held positive views on chatbot use in education, yet a nearly equal proportion expressed concerns about future impacts, especially regarding academic integrity and fairness.

The study found significant differences across demographics. Male students and those in engineering and technology fields reported higher usage and more favorable attitudes, while female students and those in humanities and healthcare expressed greater skepticism and ethical concerns. Academic level also influenced responses, with graduate students being more open to chatbot use and perceiving greater educational benefits than undergraduates. Importantly, ChatGPT usage was strongly correlated with more positive perceptions and less concern about AI’s risks.

These findings all underscore the urgent need for tailored AI policies and literacy training in higher education—ones that consider the diverse attitudes, experiences, and needs of different student populations. In my own classes, I always show students visual examples of bias in AI. My students seem to connect a little more to the visual examples and it is easy for us to then discuss other elements of written bias together more critically.

Zhang, S., Zhao, X., Zhou, T., & Kim, J. H. (2024). Do you have AI dependency? The roles of academic self-efficacy, academic stress, and performance expectations on problematic AI usage behavior. International Journal of Educational Technology in Higher Education, 21(1), 34.

Cited by 103

In a 2024 study published in the International Journal of Educational Technology in Higher Education, Zhang et al. explored a timely question: what makes students over-rely on tools like ChatGPT, and what are the consequences? Surveying 300 college students in Seoul, the researchers used the I-PACE model (a framework often used to study internet addiction) to understand how academic self-efficacy, stress, and performance expectations shape AI usage. Interestingly, while low self-efficacy alone didn’t directly predict AI overuse, it did so indirectly. Essentially, students who doubted their academic abilities felt more stressed, which led them to expect more from AI, ultimately increasing dependency.

So what does AI overuse look like? Students identified a range of negative effects, with the top five being increased laziness, restricted creativity, the spread of misinformation, and diminished critical and independent thinking. Despite AI’s academic benefits—84% used ChatGPT for help with assignments—this study urged educators to intervene.

What I stand perplexed on for this—and many other studies that continue to tie low self-efficacy, high stress, and ChatGPT usage together—is where we as teachers fall in the regards to responsibility. In many ways, teachers can accidentally act as therapists with no training for it. And perhaps a more helpful intervention for our students should focus not only on how we can reduce their stress (i.e. not giving meaningless busy work) but also how to help them advocate for less stressful environments. Many students feel the stress of school because they see it as a means to an end: I get this degree to get this job.

But what if we reframed education not as a transactional pathway, but as a space for practicing autonomy, ethical reasoning, and resilience in the face of complexity? Instead of treating stress as an individual weakness to be managed (or outsourced to AI), we might consider how institutional and pedagogical structures amplify it—and how our students can be part of reimagining those structures. When students turn to ChatGPT, they aren’t just trying to “cheat.” Often, they’re trying to survive in systems that make deep learning feel optional and mental exhaustion feel inevitable. If that’s the case, the ethical challenge isn’t just about whether students should use AI. Instead, it’s about whether the conditions we’ve created make it the most rational choice.

In that light, teaching becomes political: not just about writing better papers but about helping students write themselves into more humane futures.

This was not a coincidence of tourism. It was a summit organized by the Vatican — a formal invitation to “digital missionaries,” content creators who align their brand with Catholicism. Among them were lifestyle bloggers, fitness-minded clergy, and fashion-forward believers with Instagram bios like wife, coffee, Catholic.

And they were there for one reason: to help fix Christianity’s branding problem.

Subscribe now

“Influencing” as Institutional Strategy

Lauren Jackson’s recent piece in The New York Times captures what this moment signals: a very deliberate pivot. With many millennials disengaging from organized faith, institutions are turning to a familiar solution—influencers.

Not just priests with podcasts. Not just churches livestreaming sermons. This is different.

This is the Vatican training people to be influencers for God.

At first glance, this feels almost quaint—like a church finally catching up to 2016? But beneath the surface is something more complex. We’re not just watching a faith go digital. We’re watching influence become a morally sanctioned medium of belief.

This raises an uncomfortable question:

When faith gets filtered through the logic of the algorithm, is it still faith—or is it performance?

The Critical Media Literacy Angle

Critical media literacy (CML) encourages us to ask who is crafting messages, why they’re being crafted, and how they shape our perception. It helps us step back from the scroll and notice the systems at play—even (and especially) when they wear halos.

Let’s apply that lens here.

Who created this message, and for what purpose?
The Vatican, not individual believers, orchestrated this summit. It’s a top-down campaign to reframe Catholicism as culturally relevant, accessible, and youthful. Essentially, rebranding a global institution via TikTok and Reels.

What techniques are being used to gain attention?

• Aesthetic minimalism meets spiritual maximalism (think: “Hot Priest lifting weights to Hillsong”)

• Day-in-the-life vlogs that blend prayer with Pilates

• Oversharing framed as vulnerability: “My journey back to the Church”

This is not a condemnation. These techniques work. But for what purpose? Do people love God because their values align with the church, or because they like lifestyle being portrayed?

What gets left out of the frame?

• The theological debates that don’t translate into 15-second soundbites

• The slow, unglamorous work of community-building and spiritual growth

In a media landscape where visibility often equals legitimacy, what happens to the invisible work of faith?

When Saying “I’m Influencing You” Becomes the Influence

Here’s what I find most fascinating: Many of these creators are transparent about their mission. They say they’re digital missionaries. They know they’re influencing you. And somehow, that self-awareness is meant to defuse suspicion.

But does it?

There’s a cultural sleight of hand happening here—as if naming the power of influence neutralizes it. As if being in on the game means you're no longer being played. But transparency is not immunity. Saying “I’m trying to influence you” doesn’t make the influence less real. It just reframes it as virtuous.

This is the same logic we see in sponsored content across industries:

• “This post is in partnership with…” → Now you know! So it’s fine, right?

• “I'm not an expert, just sharing what works for me…” → So if it fails, it’s not manipulation, it's just vibes.

When the Church adopts this logic, we should pause. Because it suggests a deeper shift, not just in communication strategy, but in how institutions conceive of trust, power, and persuasion in the age of social media.

Algorithmic Faith

There’s something undeniably seductive about watching a 26-year-old woman film herself lighting a candle in a cathedral, set to lo-fi worship music, hashtagged #coffeewithJesus. It’s soft. It’s pretty. It’s emotionally resonant.

But that’s also the point.

Platform logic favors what is digestible, shareable, and easy to feel. Not necessarily what is true, complex, or grounded in theology. The Gospel of Instagram is one of relatability.

CML teaches us to ask:

• Are we engaging with belief, or with its aesthetic packaging?

• Who benefits when spiritual authority becomes entangled with personal brand?

• Can a 30-second video truly carry the weight of centuries of faith tradition?

The Church has long been a master of symbolism: incense, icons, ritual, rhythm. But what happens when those symbols become content?

Toward a More Critical Engagement

None of this is to mock the faithful. Faith, after all, has always adapted to new mediums. We can see this in the transitions from scrolls to stained glass to radio sermons. The question is not should religion evolve, but how, why, and for whom.

If you're a young person navigating belief in the age of algorithms, you deserve space to ask:

• Am I moved because of God, or because of the music, the lighting, the influencer’s glow?

• Do I feel connected, or simply emotionally primed?

• Does this deepen my practice, or just scroll past it?

The Vatican’s influencer summit marks a turning point. It asks us to confront what happens when even our most sacred institutions adopt the aesthetics of branding and the tools of social capital.

If everything is content, can anything still be sacred?

Have thoughts? Hit reply! I'd love to hear your take on digital faith, performative belief, or where you see media literacy fitting in.

Subscribe now

—Dr. Sydney

In this post, I want to share a few ways I’ve been reworking my online classes in light of generative AI. These changes aren’t about playing whack-a-mole with cheating or making my assignments “AI-resistant.” Instead, I’m re-centering my pedagogy around learning rather than credentialing—something scholars like and have long argued for. In short, I’m asking: what would happen if we stopped designing assignments with suspicion in mind and started designing for student agency and growth?

Subscribe now

Understanding Why Students Misuse AI

One of the most common concerns I hear from educators is: “My students are using AI to cheat; how do I stop it?” But that’s not the most productive question we can be asking. Instead, we should be asking: Why do students feel the need to outsource their work to AI in the first place?

A 2009 study published in the Journal of Business Ethics explored this very question. Researchers surveyed students on their motivations for cheating and found that over 60% of both business and non-business majors admitted to it. The top reason? A “desire to get ahead.” Not laziness. Not hatred of the subject matter. A sense of urgency and competition in a system that often values grades over growth.

This insight changed how I think about AI use in my classroom. If students are primarily motivated by fear of falling behind or pressure to achieve, then cracking down with harsher rules won’t solve the root problem. Instead, educators can move towards creating a learning environment where students don’t feel the need to cut corners because they’re engaged, supported, and genuinely learning.

1. Rethinking “Busy Work”: Speaking Instead of Writing

Let’s talk about discussion boards.

I’ll be honest: I’ve always had mixed feelings about them. But after the rise of AI tools, my skepticism deepened. If students can ask ChatGPT to summarize a reading and generate a post in under a minute, what are we even measuring? Worse, if the format itself doesn’t feel meaningful, students are more likely to disengage—whether they’re using AI or not.

This summer, I replaced weekly discussion posts with short “video check-ins.” Students record 5-minute reflections on their readings, share key takeaways, raise questions, and connect concepts to their lives. These videos are posted to a class Canvas discussion board where peers can comment or respond.

The impact has been huge, and I get a better sense of their authentic voices and thinking. It’s harder to fake comprehension in a video than in a text box, and as a bonus, students are building confidence with basic digital communication tools like creating a video for their peers—a critic

al skill in today’s media landscape.

2. Co-Creating AI Policies With Students

One of the most transformative shifts I’ve made is moving away from a top-down approach to academic integrity and instead co-creating classroom AI policies with students.

This summer, I invited my students to define what ethical and unethical use of AI looks like in our course. Using Padlet and small group discussions, they shared perspectives on when AI tools feel helpful (e.g., brainstorming, grammar support) and when they cross a line (e.g., generating full essays or citations without comprehension). Together, we developed a shared agreement that emphasized transparency, critical thinking, and learning over performance.

The result? Buy-in. When students have a voice in setting expectations, I believe they’re more likely to honor them. More importantly, these conversations modeled the very kind of digital literacy I hope they take into other classes and life beyond school.

(If you want the full breakdown of this model, I have a dedicated Substack post on AI policy co-creation).

3. Revisiting Flipped Classrooms

After a recent post from , I’ve also found myself revisiting the flipped classroom model—not as a “new” idea, but as a framework that aligns beautifully with the presence of AI.

In a flipped classroom, students encounter foundational content (readings, videos, lectures) before class time. Then, synchronous sessions are used for deeper analysis, peer work, or applied practice. Essentially, anything that benefits from live interaction and can’t be easily outsourced to a chatbot.

When it comes to AI, flipped classrooms offer two advantages. First, they front-load low-stakes learning, which students can supplement with AI tools in transparent ways (e.g., asking AI to explain a difficult concept or generate an outline). Second, they preserve class time for the kind of work AI can’t do well: authentic discussion, debate, collaboration, and reflection.

This model positions AI as a potential partner in the learning process, not a threat to be neutralized. It also allows me to assess learning in real-time, with assignments that privilege voice, originality, and connection.

Final Thoughts

None of this is really about “AI-proofing” in the literal sense. If a student really wants to use ChatGPT to write a paper behind my back, they’ll probably find a way to do it. But by shifting my classroom culture—from compliance to curiosity, from busy work to authentic expression—I’ve found that fewer students even want to go that route.

AI is pushing us to reconsider how we teach, but maybe that’s not a bad thing. Maybe it’s the push we’ve needed to finally let go of outdated practices, ask better questions, and build classrooms that are less about getting ahead and more about growing together.

If you’ve tried any of these approaches—or are considering them—I’d love to hear from you in the comments. How are you rethinking your classroom design in the age of AI?

That’s the question I posed in the first week of my fully online summer writing class. I’ve taught variations of first-year writing for nearly a decade, but this time, I decided to try something new: instead of enforcing a top-down AI use policy, I invited my students to co-create one. Something I was sincerely nervous about.

Butt it worked better than I imagined.

Subscribe now

In a time of mounting anxiety around academic integrity, AI misuse, and student disengagement, this simple activity flipped the script. Not only did students take the assignment seriously, but they also articulated a thoughtful, nuanced understanding of how AI can support (or sabotage) their learning.

In this post, I’ll share exactly how I structured the assignment, what students said, the policy we built together, and why I believe this model is one of the most promising ways to cultivate critical digital literacy, ethical engagement, and student agency.

Why I Let My Students Write the Rules

This wasn’t just a warm-and-fuzzy attempt at classroom democracy. The decision to co-create our AI policy was rooted in three pedagogical principles:

1. Contract grading: Our course is built on Asao Inoue’s labor-based grading contract model that emphasizes effort, reflection, and growth—not perfection. This makes space for risk-taking and collaboration, rather than competition for points.

2. Critical digital literacy: As AI tools like ChatGPT become more integrated into students’ daily lives, we can’t afford to teach around them. We need to teach through them—focusing on ethics, impact, and the habits of mind that will outlast any one platform. Much like critical media literacy scholars have been teaching for decades.

3. Community agreements over surveillance: Rather than treating students as cheaters-in-waiting, I wanted to approach policymaking as a collaborative act of trust, one that aligned with our course values of compassion, equity, curiosity, and care.

So I ditched the lecture about “AI misuse” and launched week one with an open invitation.

The Assignment: Co-Creating an AI Policy

Here’s what I posted on Canvas to kick things off:

“Imagine you’re in a home studio writing group—not a university classroom. You're not trying to earn a grade, but to grow as a thinker, writer, and community member. In this space, we use tools, learn from each other, and figure things out together. One of those tools might be AI—but what should that look like?”

Students were directed to post their thoughts on a shared Padlet board and reply to at least two peers. I gave them a range of response options: short paragraphs, guidelines, questions, do/don’t lists, etc.

They also had guiding prompts like:

• What types of AI use feel ethical, helpful, or appropriate in a writing course

• When does AI help you learn—and when might it hinder your growth?

• How can we ensure AI use doesn’t impact our critical thinking skills?

Before they participated, students were required to read our full grading contract (based off Inoue’s fully accessible Grading Contract Template), which frames the classroom as a cooperative learning environment grounded in trust and mutual responsibility. This helped set the tone: their words would help shape real policy, not just check a box for participation.

What Students Said

Over the course of the first week, dozens of thoughtful, surprising, and even funny responses rolled in. Some paraphrases:

• “AI can be a tool, not a crutch. If it does the hard part for you, you won’t grow.”

• “It’s like using Grammarly but on steroids. I think it’s okay as long as you’re still learning from it.”

• “What’s the point of taking a writing class if you’re not going to try writing?”

• “AI is helpful for outlines and ideas, but not for actual voice.”

• “We need to be transparent. If you used AI a lot, say so.”

Several students admitted they’d experimented with AI before but hadn’t felt comfortable asking what “counted” as cheating. Others expressed relief that we were even having this conversation. Across responses, a clear agreement emerged:

General Agreement:

AI can be a valuable support tool for learning when used intentionally and ethically. It should assist—not replace—your thinking, voice, or writing process.

The Policy We Wrote Together

Based on their responses, I compiled and shared our first version of a Class-Designed AI Policy. Here's a condensed version:

Ethical Uses of AI:

• Brainstorming ideas or generating writing prompt

  • Example: Asking AI to help come up with hooks or thesis starters.

• Clarifying confusing concepts or assignments

  • Example: Using AI to simplify a reading or explain a complex term.

• Grammar and sentence-level revision

  • Example: Asking AI to rephrase an awkward sentence or check for run-ons.

• Organizing ideas or creating outlines

  • Example: Getting suggestions for how to structure your argument.

• Feedback and growth-oriented suggestions

  • Example: Asking AI to evaluate clarity or tone and suggest improvements.

➤ Inappropriate Uses of AI

• Writing full essays or paragraphs

• Answering quizzes or graded questions

• Replacing your original voice

• Using AI heavily without disclosing it

➤ Transparency Expectation

“If you use AI in a significant way, include a short note in your author’s memo or margin comments.”
Example: “I used ChatGPT to help brainstorm my topic and reword the intro paragraph.”

➤ Why This Matters

Students emphasized that learning requires effort. If AI does all the work, they lose out. Voice, growth, and originality were recurring values.

➤ Transparency and Opportunities for Revisions

At the end of our AI document I added the following:

“You can find the full version of the discussion here (linked out to PDF exported Padlet, not included in this Substack for student privacy). If for some reason you feel your views were not represented in this official document, please feel free to reach out to me at [school email]. This policy was designed to represent a wide range of ideas while still maintaining a general consensus around AI and its uses in this particular class.”

Why This Works

While I almost assumed this activity would be nothing more than a “feel good” exercise, I was delighted to see how deeply practical it was. Here’s why I’ll keep doing it:

1. It builds buy-in (I hope).

I don’t have any data to confirm this—yet—but I hope students are more likely to follow this AI policy because they helped write it.

2. It invites ethical reasoning.

Instead of memorizing rules, students practiced critical thinking, negotiation, and values clarification—essential skills for digital literacy.

3. It reframes “academic integrity.”

Rather than fear-based language or surveillance, this approach fosters responsibility and reflection. It sends the message: we trust you to think about your choices, not just follow commands.

4. It humanizes AI.

I did find that students were more curious—and less anxious—about AI tools after this conversation. We created space to explore, not just comply.

What I’d Change Next Time

• Incorporate examples of gray areas: Students wanted to know more about “in-between” cases, like using AI to reword a single sentence or help with citations.

• Bring in student voices from other classes or schools as comparison points (maybe via anonymous quotes or guest posts).

• Track changes over time: I’d love to revisit the Padlet mid-semester to see if their views shift.

Try This in Your Own Classroom

Want to run your own version of this activity? Here’s a plug-and-play version:

Quick Start Guide:

1. Set the tone with a message about learning, trust, and shared responsibility.

2. Use a collaborative tool (Padlet or Google Doc) for students to post.

3. Offer flexible formats—lists or questions.

4. Provide guiding prompts (feel free to steal mine above).

5. Compile responses into a living policy and revisit it throughout the term.

6. Hyperlink the policy in your syllabus or LMS and give students credit for their work.

Let me know if you'd like a downloadable version of the template I used or a sample Padlet prompt. I’m happy to share!

Final Thoughts: Policy as Pedagogy

This week on LinkedIn I saw a post from Maurie Beasley, M.Ed that read:

Ultimately, this AI-policy activity wasn’t about AI. It was about values. When we invite students to co-create the boundaries of our classroom, we teach them something far more important than citation rules: that they are ethical agents and critical thinkers, capable of shaping the culture of learning we build together.

As AI evolves, our policies will, too. But what should remain constant is our commitment to shared dialogue, student voice, and the collective work of thinking through these tools—not just policing them.

I’d love to hear from you. Are you thinking about co-writing policies with students this fall? How are you approaching AI in your classroom?

Let’s keep the conversation going.

—

Bonus: Want a template for your syllabus or LMS? Drop a comment or reply and I’ll send you a copy of the full Padlet prompt + class policy document I used.

I knew The Materialist was being marketed to me before I even saw a trailer. One by one, the podcasts that punctuate my commute—each with its own vibe and politics—started gushing about the film. The praise never felt scripted, yet the phrasing was uncannily consistent: “smart satire,” “can’t-miss commentary on consumer culture,” “mind-bending third act.” Four hosts reading from the same invisible brief. All with a similarly vague sponsorship disclosure: “This episode may feature content that has been externally sponsored or internally incentivized.” Obviously done so the disclosure was there but what was being sponsored was up for interpretation. One thing was clear to me though…My feed had become a synchronized chorus.

Influencer marketing has always relied on trust, but today’s campaigns are engineered for total saturation. Studios buy share of voice—not just ads, but talking heads, TikTok skits, pop-culture newsletters, pre-release “critic” tweets. The goal is to replace serendipitous buzz with manufactured inevitability: of course you’ll watch The Materialist; everyone agrees it’s great.

Subscribe now

Enter the Algorithmic Rebuttal

Yet almost as quickly, a countercurrent has formed. Instagram critiques say the movie supports “the broke man propaganda.”

TikTok critics remark the movie as “a flop.” Longform Substack essays dissected plot holes with surgical precision. Instead of a consensus, the internet produced a whiplash oscillation—5-star raves versus visceral takedowns.

And this is the kind of critique I love to see. Something that goes against the echo chamber of enthusiasm and asks deeper, more critical questions. To me, these critiques are not just reactions, but a form of digital protest. When influencers flood our feeds with glowing praise that mirrors each other word for word, it’s easy to feel like our own thoughts don’t matter. That’s why so many people flock to the comments, stitch videos, or post takedowns: it’s not just to be contrary—it’s to be real. To say, “I see through this,” or “this didn’t speak to me,” or “stop pretending this is organic.” These are moments of resistance in a culture that increasingly asks us to perform enthusiasm on cue. And while, yes, that backlash still keeps the movie in the spotlight, it also carves out space for people to express their values, their standards, and their desire for authenticity—on their own terms.

Critique can be a form of collective immune response. Our digital antibodies are being alerted to hidden persuasion. When TikTok creators reverse-engineer sponsorship disclosures or when podcasters openly compare talking points, they perform media literacy in real time. The question is whether this resistance scales beyond performative negation. Does yelling “ad!” actually inoculate us, or does it just feed the content mill with another monetizable posture?

True critical agency would mean holding several thoughts at once: yes, influencers are paid; yes, algorithms amplify dissent; and yes, a film can be enjoyable even under those conditions. It would mean carving out space for first-hand experience—watching The Materialist (or not) on our own terms, then reflecting before broadcasting hot takes.

Toward a Pedagogy of Algorithmic Taste

This tension—between manipulated consensus and contrarianism—is fertile ground for classroom exploration. Media literacy often stops at decoding advertising techniques; today we must also interrogate platform logics that shape not only what messages reach us but how swiftly counter-messages appear. Teaching students to map the feedback loops of hype and backlash equips them to notice when their adoration (or sometimes contrarianism) is being instrumented.

Turning the Think Piece into a Classroom Lesson (College Writing or Media Studies, 1–2 class periods)

1. Warm-Up (15 min)

Show a montage: 3 influencer clips praising The Materialist, followed by 3 viral negative reactions. Quick write: “What do these voices want from you?” Surfaces gut reactions and primes discussion about persuasion.

2. Mini-Lecture & Discussion (20 min)

Briefly define: influencer marketing, algorithmic amplification, and backlash culture. Use real data.

3. Evidence Hunt (30 min)

In small groups, students track one talking point across platforms (podcast, Instagram, Reddit) and document similarities/differences. Practices source comparison and pattern recognition (i.e. recognizing that 5 different podcasts have talked about The Materialist within the same week and have similar praise).

4. Reflection & Position Draft (20 min)

Individually, students draft a 200-word micro-review of The Materialist imagining they have not seen it. Student then annotate which parts are influenced by external narratives. Highlights how pre-exposure frames expectation.

5. Homework / Extension

Watch the film (or selected scenes) and revise the review after viewing. Add a meta-paragraph: “What changed after firsthand experience?” Reinforces distinction between mediated expectation and personal judgment.

6. Assessment

Short essay or discussion post analyzing the “purpose” of advertising The Materialist. Student can also answer whether online backlash functions as meaningful resistance or just another commodified cycle. Checks critical synthesis & application.

Optional Tech Tie-In: Use a tool like Padlet to visualize engagement spikes for positive vs. negative keywords around release week. Students can plot correlations between marketing pushes and backlash peaks.

Closing Thought

Advertising will keep hiding in plain sight, and the antidote isn’t abstinence from media but deliberate pauses. Where are the moments where we notice the choreography and ask, “Who benefits from my reaction?” Teaching those pauses, whether through college writing prompts or K-12 media classes, might be our best shot at cultivating viewers who can love—or hate—a movie on their own terms.

I’ve been a big believer since AI came out that policing AI won’t help students become more critical or ethical thinkers. However, what AI-integration looks like in practice becomes much murkier. Lately, I’ve been thinking about what it would look like to take a cue from labor-based contract grading models like Asao Inoue’, and co-write our classroom AI policies with students. Here’s are my thoughts on what this could look like.

If you're new to contract grading, here's a quick summary: Instead of assigning grades based on a teacher's assessment of quality, Inoue’s labor-based contract grading emphasizes the student’s effort, process, and engagement. It asks: what happens when progress is measured by commitment rather than perfection? This model fosters equity by honoring diverse forms of expression and valuing the work students put in—especially those from marginalized backgrounds. You can read more about Inoue’s approach and some practical resources here.

So, what does this have to do with AI? A lot.

Subscribe now

From Policing to Partnership

When we co-write AI policies with students, we shift the classroom from a place of surveillance to a site of inquiry. Students learn not just whether to use AI, but why, when, and how. This can also mean discussing explicitly with students what it means to use AI responsibly.

Like Inoue’s model, this approach centers student labor and rhetorical decision-making. We’re not grading "AI-free" work or punishing "AI-sounding" essays. Instead, we’re asking students to document their process, analyze what AI generates, and reflect on the choices they made. In doing so, students can build metacognition and digital literacy in tandem.

Addressing Bias and Burnout

AI tools aren’t neutral. They replicate dominant norms and suppress linguistic diversity—just like traditional grading systems often do. When students co-create classroom policies, they’re better equipped to interrogate these tools' biases and limitations. This practice can also reduce burnout: it transforms AI from a looming threat into a shared challenge we can tackle together.

Start With the School Policy

Before inviting students into this conversation, make sure you’ve reviewed your school or district’s official AI policy. If one exists, be transparent about its parameters. If one doesn’t, you have the chance to model ethical leadership by building a policy with your students from the ground up. Just be sure to note that your classroom policy must align with or build on any institutional guidelines in place.

The Process: Co-Writing with Care

1. Begin with Inquiry. Ask students: How are you already using AI? What do you think is fair, ethical, or helpful?

2. Build a Foundation. Most students don’t know where AI came from, or what it can do. I would recommend having students watch a comprehensive video such as AI and the Future of Us, which explores impact of artificial intelligence on people's daily lives with several leading researchers and developers.

3. Discuss Ethical Frameworks. Talk about bias, hallucination, and misinformation. All of these examples can help students understand why we cannot just trust AI output “as is” and need to do more critical thinking.

4. Analyze Examples. Provide AI-generated writing and invite critique. What would you keep? What needs human judgment?

5. Draft Together. Use a shared doc, whiteboard, or Padlet to outline your policy. Include expectations, boundaries, and revision protocols.

6. Revisit Often. AI is evolving. Your policy should, too. I would recommend using your first student draft as a “base” but allowing each new class to review and approve this draft. However, you should keep the policy in place during the entirety of a semester or a quarter to avoid confusion with grading or AI policy.

Co-creating an AI policy rooted in equity isn’t about letting students use ChatGPT unchecked. It’s about equipping them to navigate a digital world full of gray areas—and helping them trust their own judgment within it.

Like contract grading, it’s less about the tool and more about the trust.

Share

Brown’s research on wholehearted living shows that creativity, play, and rest are not luxuries but “essential to innovation.” Yet the gateway to all three is often boredom. It’s those uneasy intervals when the mind wanders, daydreams, and stitches disparate thoughts into something novel. In Atlas of the Heart, Brown defines boredom as “the uncomfortable state of wanting to engage in a satisfying activity but being unable to do it,” a feeling that can wind us up with irritation or launch us into imaginative flight. Neuroscientists back her up: when external stimuli drop, the brain’s default-mode network lights up, connecting seemingly unrelated memories and concepts that become the raw materials of innovation. Children, especially, need that mental white space; without it, they can only remix what someone—or something—else has already made.

Subscribe now

Psychologists have quantified the effect. In a classic experiment, participants who first spent fifteen minutes copying numbers from a phone book, a task designed to induce tedium, later outperformed a control group on idea-generation tests, dreaming up more novel uses for paper cups and bricks. Researcher Dr. Sandi Mann explains that monotony encourages the mind to “look inwards,” activating daydreaming processes that stitch together remote associations. Brown’s framework dovetails neatly: boredom is the pause that lets vulnerability and imagination shake hands. If we erase that pause, we don’t just lose a pleasant diversion; we lose a proven cognitive catalyst.

I heard about the Mattel’s partnership with AI during my co-hosted podcast The Disinfo Detox. Our resident philosopher, Dickey, introduced and I to Mattel’s plans to partner with OpenAI and ship its first generative-AI product by year-end, promising Barbie dolls and Hot Wheels tracks that talk back, compose stories, and personalize play experiences in real time.

After the podcast ended, I took a deep dive into toys and OpenAI. Amazon is rolling out Alexa+ features like “Explore with Alexa,” turning the family smart speaker into a homework helper that spins adaptive trivia adventures. Start-ups are pitching plush companions such as Curio’s Grok, while Yukai Engineering’s shy, sensor-laden Mirumi clings to backpacks, and the educational robot Miko 3 boasts Disney narratives and on-demand chess lessons. The sales pitch is identical: no more idle moments, because the toy will always have something “smart” to say.

There is undeniable wonder in these inventions. Isn’t it miraculous to have a talking train track can scaffold vocabulary? Or an AI plushie can comfort a child who finds human peers overwhelming? OpenAI operating chief Brad Lightcap stated that “With OpenAI, Mattel has access to an advanced set of AI capabilities alongside new tools to enable productivity, creativity, and company-wide transformation at scale." And this seems hopeful. But developmental psychologists warn that children thrive most when they shape their own narratives, not when a device scripts them.

Brown reminds us, “There is no innovation and creativity without failure. Period.” However, failure is pretty impossible if your digital playmate politely corrects every mispronounced word or edits your story arc on the fly. The risk is not that AI toys will make children less intelligent. It is very possible that AI may boost some skills. The fear is that AI toys will make the imaginative terrain too smooth to stumble on. A generation that never builds a cardboard spaceship because a plastic one already narrates a perfect journey could grow up excellent at refinement and weak at invention, feeding the same algorithmic loops that power the toys in the first place.

In media studies the phenomenon is called “regurgitative culture.” The idea is that machines remix the recent past until everything looks uncannily familiar. When children outsource pretend play to a data-trained character, they receive stories optimized for retention metrics, not surprise. A Hot Wheels track that auto-generates commentary based on last season’s race footage might entertain, but it also narrows the possibility space. Tonight’s chat logs feed tomorrow’s model weights, which script next year’s toys. With every cycle the palette of ideas gets a shade more uniform, and the society that emerges risks running on reruns.

None of this is an argument for smashing the circuit boards. AI companions can augment accessibility for neurodivergent kids, translate play into multiple languages, and expose rural classrooms to STEM concepts that would otherwise remain abstract. The challenge is calibration. Brown encourages parents to “let go of exhaustion as a status symbol,” and we might extend the mantra to stimulation: we can choose not to valorize constant novelty. That means carving out deliberate tech-free intervals/ It means Sunday afternoon downtime, the walk to school, even the boring wait at the DMV. All so that children’s minds have room to wander beyond scripted conversation trees.

The next time your child sighs, “I’m bored,” consider it a creative invitation rather than a crisis. Resist the reflex to hand over the newest talking doll or ask Alexa to queue another game. Instead, sit with the discomfort for a beat; let the silence stretch like taffy. Even if nothing visibly happens, beneath the surface the brain is lighting up like a switchboard, rehearsing possibilities. Those minutes may feel unproductive, but, as Brown’s work suggests, they’re where originality germinates. If we can protect boredom in an era obsessed with frictionless experience—if we can champion the blank page alongside the smart screen—we won’t just raise better innovators; we’ll raise kids who trust their own imaginations enough to build a future none of our algorithms can predict.

Subscribe now

Thanks for reading. If this resonated, share it with someone else who believes in the power of a little boredom.

Warmly,

Dr. Sydney

“While it is utterly unreasonable to expect classrooms to be therapy sessions, it is appropriate for [students] to hope that the knowledge received in these settings will enrich and enhance them.”

-bell hooks

Lately, I’ve been turning this quote over in my head as I scroll through the latest updates in AI, read research on student well-being, and sit in front of my summer syllabus wondering: What do I even do with this?

A few years ago, I might have confidently told students how to navigate the world of information, how to compose a thoughtful argument, how to spot misinformation. But now—when AI tools can draft a persuasive letter in seconds, replicate citation styles, even generate entirely falsified videos—my role feels newly uncertain. Not because I don’t still believe in the importance of those skills, but because the conditions surrounding them have changed so rapidly that it feels like we’ve all been dropped into an entirely different classroom. One with new rules, new ethical tensions, and no clear historical precedent.1

What’s most disorienting is that this is happening in real time. Students aren’t the only ones encountering AI for the first time—educators are too. The shift didn’t arrive in our rearview mirror with neat conclusions and long-standing pedagogical debates. It showed up at the front door, took a seat in the classroom, and started completing our assignments before we could finish explaining them.

We’ve spent so much of our teaching lives operating under the assumption that we’re preparing students for what’s next. But AI has flipped that. We’re not leading them forward into clarity—we’re standing with them inside the fog.

The real tension lies in the fact that educators are still expected to teach with authority, to assign meaning, to help students make connections between what they’re learning and who they’re becoming. But with AI, we don’t always know what that connection is yet. We’re watching our disciplines reconfigure themselves. The “life experience” we’re supposed to tether students’ learning to is still unfolding for all of us.

Teachers have always responded to cultural shifts. Whether it was the arrival of television, the first desktop computers, or the rise of the internet, teachers have been there to respond and adjust. But AI brings a new kind of disorientation: one where the pace of change is exponential, the ethical ground is unstable, and even the idea of authorship feels up for grabs. And we’re not guiding students through it from a place of hindsight; we’re learning it as they do.

We’re all feeling it. There is a weight of knowing that our students are looking to us not just for information, but for meaning—and realizing that we don’t always have it yet. That can lead to guilt. Impostor syndrome. Or the quiet, gnawing sense that we’re not doing our jobs well enough.

Research indicates that teachers are transitioning from being mere content deliverers to facilitators of learning, guiding students in navigating AI tools effectively. Zhai (2024) proposes a framework categorizing teachers' engagement with AI into four roles: Observer, Adopter, Collaborator, and Innovator. This model underscores the importance of teachers evolving alongside technological advancements to maintain their relevance and efficacy in the classroom.

Moreover, the emotional and ethical dimensions of teaching are gaining prominence. As AI handles more administrative and instructional tasks, the human aspects of teaching—such as empathy, ethical judgment, and the ability to inspire—become even more critical. Educators are now tasked with fostering environments where students can critically assess AI's role and implications in society.

In this context, it's essential for teachers to embrace continuous learning and adaptability. Engaging in reflective practices allows educators to assess their evolving roles and develop strategies to integrate AI in ways that enhance, rather than diminish, the educational experience. By doing so, teachers can model lifelong learning and critical thinking, equipping students to navigate an increasingly AI-integrated world. Perhaps it’s time for us to help students pause just long enough to ask: What does this technology mean for us as people?

AI doesn’t just challenge our assignments; it challenges our identities—our sense of agency, creativity, even purpose. And if that’s disorienting for us as educators, imagine how it feels for students who are still forming those identities in the first place.

This is why our work still matters. Perhaps now more than ever. Not because we have all the answers, but because we can still help students navigate the emotional and ethical terrain of living with technology that moves faster than human development.

So no, the classroom isn’t therapy. But it can be a space for reflection, for meaning-making, for learning how to live wisely in a world that often confuses speed with wisdom.

Practical Educator Takeaways for Transgressing with AI

1. Narrate your own learning.
When introducing AI tools or assignments, talk openly about what you are still figuring out. Students don’t need us to be experts (or even see us as experts) they just need us to be honest and thoughtful as we make sense of emerging tools.

• In Teaching to Transgress bell hooks encourages teachers to “take the first risk” when it comes to linking confessional narratives to academic discourse. We can start making these confessions about AI with our students.

2. Create space for student meaning-making.
Design assignments that ask students to reflect on how AI tools intersect with their own values, goals, and experiences. Reflection journals, media analysis, and ethical debates can deepen the personal connection to course content.

3. Reframe “academic dishonesty” as a teachable moment.
Rather than rushing to punish AI misuse, use it as an opportunity to explore why students might be turning to these tools. Is it overwhelm? Confusion? Lack of confidence? These insights can guide deeper conversations around trust, labor, and learning.

4. Slow down.
It’s okay not to overhaul your entire course in one semester. Introduce changes incrementally. Give students (and yourself) room to breathe, try, fail, and iterate.

5. Co-create AI policies with students.
Invite students into conversations about how, when, and why AI might be used in your class. Building guidelines together can foster transparency and shared accountability.

• Asao Inoue’s work on Contract Grading may be helpful in thinking about student-centered curricula.

6. Connect with colleagues.
You’re not alone. Especially in spaces like Substack where hundreds if not thousands of authors are writing, researching and collaborating on AI. Whether through informal discussions or formal PD, carve out time to process what’s happening with others. Here are some good people and places to start:

7. Remember your purpose.
You don’t need to be an AI expert to teach with purpose. If your classroom helps students grow in curiosity, integrity, and compassion, you’re doing the work that matters most.