Agents, Agency, and Action...

Academic Year 5 of AI is less than 3 months away...

Year 5 Needs to Be the Year of Curriculum Change

In the first piece of this series where I reflect on some common themes and ideas that emerge from the last 6 months of talks, workshops, panels, and consulting, I tried to sit with militant apathy and the artifices of education. In the second, I tried to name the audacious ask of higher education: the degree to which we ask students to invest money, time, energy, attention, and hope into a system that offers no guarantees and often limited agency over what the experience actually looks like.

Those two things sit together for me because they point to the same problem. Students are responding to higher education’s contradictions being made more visible by AI.

And so, as we move toward the Fall and what will be academic year 5 of GenAI (yes, you read that right, year 5) being widely available, the thing I keep coming back to is this: Year 5 needs to be the year of curriculum change.

We have had time for syllabi statements, policies, how to use AI, how to not use AI, moving beyond the academic integrity conversation, developing AI literacy frameworks and saying that there’s not enough time. We’ve spent nearly 4 years in this conversation, and while these are still important, at year 5, they are not enough.

Students are now graduating from higher education having had four years of GenAI existing in the world and, for many of them, their experience of it in college has been scattershot. Maybe one faculty member talked about it. Maybe another banned it. Maybe one assignment required it. Maybe another treated its use as misconduct. Maybe one course helped them think critically about it. Maybe another never named it. Maybe one faculty member gave specific guidance, and another said “don’t use AI” without defining what that meant.

But our students need and deserve more than curriculum roulette; another room in a house they’ve bought that they cannot actually see (for that reference, see part 2 of this series).

If we are asking students to take on debt, to invest years of their lives, to trust that this thing called education will help them move into a future that none of us can fully see, then we cannot also leave their engagement with one of the most disruptive technologies of the moment to chance, personality, individual comfort, or whoever they happen to get for a Tuesday/Thursday section at 9:30am.

That does not mean every course needs to become an AI course where it centers AI, talks about AI every week, uses AI in every assignment, or becomes some weird half-course, half-tech-demo situation. No. One. Wants. That…except maybe the techbro billionaires trillionaires.

But every program does need to ask what AI means for its curriculum. Every department needs to ask what students need to understand, practice, resist, use, critique, avoid, and evaluate in relation to AI in that field. Every general education program needs to ask what it means to prepare students for a world where AI is simultaneously present in work, media, research, writing, communication, administration, hiring, creativity, surveillance, and all kinds of decision-making systems.

For the last few years, a lot of higher education has been in reaction mode. ChatGPT showed up publicly in late 2022, and many of us were trying to understand what it was while simultaneously being told what it was going to become by people who had very strong incentives to make it sound inevitable, magical, catastrophic, or all three, depending on the pitch deck. Meanwhile, its availability, affordances, and affordability continued to change and shift, making it hard to hold long enough to have a clear understanding of what we are even talking about when we talk about “AI.”

We were also not exactly working in a calm moment (I mean, for the past 10 years, I’m not sure one has existeed). Higher education was already exhausted from the pandemic, already navigating enrollment pressures, public distrust, political hostility, labor challenges, financial constraints, burnout, and a general sense that someone keeps pouring gasoline on the dumpster fire.

This scattershot response was to be expected. Just because it makes sense, it doesn’t mean it can continue. At some point, the curriculum has to change.

That does not mean the curriculum has to become settled. In fact, I think one of the hard things here is that we have to build a curriculum around something that will not sit still. Things AI could not do three years ago, it can do now. Things it did poorly a year ago, it may do passably or even well now. Things it does well in one context, it may still do badly in another. Hallucinations matter differently depending on the model, the task, the tool, the user, the stakes, and whether anyone knows enough to check the output.

Therefore, the curriculum cannot just be “teach students the tool.” That will age badly before the next catalog revision. The curriculum ask is disciplinary judgment training.

How do students learn to judge when AI is useful, when it is harmful, when it is flattening their thinking, when it is extending their capacity, when it is producing nonsense, when it is giving them a false sense of understanding, when it is helping them access something they were otherwise blocked from, when it is replacing the very practice they needed to struggle through, and when the refusal to use it is itself an important ethical or intellectual choice? There are no single clear answers; all need to be grounded in developed judgment.

And the answer to the curriculum cannot be to entirely ban AI. That would be a discipline’s utter failure to grapple with change in a productive way. The kind of judgment students need to develop cannot be built if students only encounter AI as a prohibition, panic, or personal preference.

Again, this doesn’t mean AI should be a free-for-all where everyone does whatever and calls it innovation. But if all we give students is rules, we have not taught ethics. Ethics is not just being prevented from making a choice, being told not to, or being judged for doing so. Ethics is learning how to make choices, how to explain those choices, how to be accountable for those choices, and how to understand the consequences of those choices for oneself and others.

That is one of the limits of policing and control. They can tell students what is and is not allowed. They can protect some boundaries. They can create some structure. But they do not solve for pedagogy. They do not solve for ethics. They do not solve for workforce preparation. They do not solve for meaning-making.

All of those require judgment, design, and intentional engagement. And this is where the curriculum conversation has to get much more concrete. The curriculum conversation should start to reassess the frictions, fictions, and feasibilities of AI.

Learning needs friction: effort, repetition, confusion, attention, and practice. But that friction may come to look different and we have to be ready to negotiate and be open to that difference.

Some frictions are just institutional nonsense or can feel that way, just like we all experience it. I have yet to meet a single person who takes IT and HR mandatory courses serious but view it as an annoyinghurdle to get through. They’re incredibly important topics and also, the trainings feel so disconnected and pro-forma that we collectively smash the “Next” button as fast as we can. And that can be also how our curriculum is experienced.

If students experience every friction as the same, then AI becomes a reasonable tool for reducing all of it. The hoop they jump through because hoops are everywhere in the institution.

We have to recognize that students feel that way about many parts of the education regime and we have to build different bridges to them than what has worked before. We have to get much clearer about which frictions serve learning and which frictions serve the institution’s need to move students through the system.

By fictions, I mean the stories higher education tells itself that AI is making harder to sustain.

One fiction is that students know why the work matters. Many students do not. That does not mean they do not care. It means we have often asked them to infer the value from the work itself, the grade attached to it, or the vague promise that it will matter later. “This will matter later” feels increasingly thin when the later students are being pointed toward is also being actively reorganized by AI.

Students can handle difference of AI uses across their course so long as it is meaingfully and clearly grounded. It is not new that disciplines, faculty, and courses have differences. The issue is coherence. Students need help understanding why those differences exist and what those differences mean, which is program-level work.

Another fiction is that this can be solved individually.

Individual faculty matter enormously. Much of the best work has started with individual faculty trying things, sharing things, failing in public, revising, asking students what happened, and bringing that learning back to colleagues. I have learned so much from faculty who have tried something with AI and then described exactly what happened,where it went sideways, where it worked better than expected, and what they would change the next time. Those details are gold for all of us and yet, it is insufficient as a system.

When AI engagement depends entirely on individual faculty comfort, curiosity, anxiety, or exhaustion, students experience the curriculum as random.

At the curriculum level, departments should be asking a set of recurring questions focused on what the discipline has figured out. I see three big questions and then many offshoots from each of those.

How is generative AI and agentic AI showing up in the field?

• How are professionals actually using it in the field?

• How are people discussing, documenting, or disclosing their use in the field?

• What are the declared limits in the field?

• Where is AI being treated as normal practice, where is it being resisted, and where is the field still pretending it is somehow outside the conversation?

What is the full discourse of generative AI and agentic AI in the field?

• What are professional organizations saying?

• What are accrediting bodies saying?

• What are employers saying?

• What are researchers saying?

• What are critics saying?

• Where does the discourse align with actual use in the field, and where does it diverge?

What does that mean for how and what we teaching in the discipline regarding AI?

• Where should students first encounter AI in the major?

• Where should they learn appropriate uses?

• Where should they work with it? W

• Where should they work without it?

• Where should they practice refusal?

• Where should they learn verification?

• Where should they learn disclosure?

• Where should they see the limits of AI in the field?

• Where should they develop judgment about when these tools help, when they distort, and when they get in the way of the very thing they need to learn?

This work also has to be feasible. Faculty are already carrying too much. Year 5 cannot become another version of higher education institutions saying, “Please solve this massive structural problem in your spare time.”

Higher education loves to identify a structural problem and then assign it to individual faculty as a matter of passion, care, or innovation. That burns through people’s good will and turns vocational awe into exhaustion or even disdain. It relies on the same labor patterns that have made so much of this work difficult already.

A curriculum ask requires institutional support: time, compensation, department-level space, student input, shared language and administrative support and follow through. A curriculum conversation without time and support becomes another email with a spreadsheet attached.

The work can still start small. A department can identify two or three places in the curriculum where students should explicitly engage AI. A program can decide where students should learn disclosure and verification practices. A general education program can decide what every student should understand about AI, information, bias, labor, creativity, and human judgment. A faculty group can revise one shared assignment and compare what happened across sections.

That will not fix everything. Still, it starts to replace curriculum roulette with design.

And while we do this work, we need students. They are already using AI and navigating the contradiction between classroom rules and the rest of the world. They have insights we need. And, they also have gaps, misunderstandings, overconfidence, and anxieties, which is why they belong in the conversation. If they are not already involved in the institution’s curriculum plans, then they need to be brought it. We can’t repeat the past where students enter the conversation last and largely just to be told what has been decided.

Year 5 should be the year we start connecting the rooms.

That means moving from scattered classroom experiments to program-level judgment. It means asking where AI belongs, where it disrupts, where it clarifies, where it distorts, and where it forces us to admit that the assignment, policy, or program structure already needed attention. From assignment level clarity to disciplinary judgment to student agency, our conversations need to be honest, intentional, and mark progress towards what the new practices and moves will be.

We also need focus. AI can represent myriad concerns and anxieties. In the department space, the focus should be on the curriculum, making sure the degree or program represents what it is suppose to achieve. That means some of the problems and concerns of AI, while still relevant, do not need to be the focus of department discussions.

We need to accept that this won’t be perfect, but higher education hasn’t never been perfect. What it can’t be is stagnant and a curriculum that hasn’t been revisited in 5 years, especially in light of a technology like this, is quickly stagnant.

That is the curriculum ask.

If apathy is about student resistance and agency to dubious claims by the institution, and if the audacious ask of higher education is only getting harder to defend, then curriculum becomes one of the places where we can meaningfully respond.

We can work with students to consider what work matters and why. Collectively, we can determine where AI fits and where it fails. We can help them see what our disciplines know that the tool cannot know on its own. We can practice judgment with them in a world that keeps offering shortcuts.

And we can do that by making the curriculum worthy of the trust we ask students to place in it.

That feels like the work of Year 5.

---

The Update Space

Upcoming Sightings & Shenanigans

• Keynote speaker at the Reimagining the Liberal Arts in the Age of AI Conference, July 21-23 at the University of Mary Washington.

• EDUCAUSE Online Program: Teaching with AI. Virtual. Facilitating sessions: ongoing

Recent Recordings, Resources, & Writings:

• Davis, L., & Eaton, L. (May 2026). Expanding OER with GenAI. EDUCAUSE Review.

• AI x Higher Ed Podcast with Anand Rao & Stefan Bauschard. Episode: Universities Must Adapt to AI—Here’s How They’re Doing It (May, 2026)

• Damm, C., & Eaton, L. (2026, March). From prompt to practice: A framework for transparent GenAI use in higher education. EDUCAUSE Review.

• Eaton, L., Nemeroff, A., & Sun, X. (2026). AI-assisted course design and development. In K. S. Ives, M. Cini, & R. Schroeder (Eds.), AI applications in online higher education administration: Strategies for maximizing returns and improving outcomes. Routledge.

• Margin of Thought with Priten: Season 1, Episode 5: How Can We Center Pedagogy During the AI Tech Wave? (February 2026)

• Online Learning in the Second Half with John Nash and Jason Johnston: EP 39 - The Higher Ed AI Solution: Good Pedagogy (January 2026)

• The Peer Review Podcast with Sarah Bunin Benor and Mira Sucharov: Authentic Assessment: Co-Creating AI Policies with Students (December 2025)

• David Bachman interviewed me on his Substack, Entropy Bonus (November 2025)

• The AI Diatribe Podcast with Jason Low (November): Episode 17: Can Universities Keep Pace With AI?

• The Opposite of Cheating Podcast with Dr. Tricia Bertram Gallant (October 2025): Season 2, Episode 31.

• The Learning Stack Podcast with Thomas Thompson (August 2025). “(i)nnovations, AI, Pirates, and Access”.

• Intentional Teaching Podcast with Derek Bruff (August 2025). Episode 73: Study Hall with Lance Eaton, Michelle D. Miller, and David Nelson.

• Dissertation: Elbow Patches To Eye Patches: A Phenomenographic Study Of Scholarly Practices, Research Literature Access, And Academic Piracy

• AI Syllabi Policy Repository: 200+ policies (always looking for more- submit your AI syllabus policy here)

Finally, if you are doing interesting things with AI in your higher ed classrooms, consider being interviewed for this Substack or even contributing. Complete this form, and I’ll get back to you soon!

We periodically host small-group workshops and leadership sessions for higher ed teams. You can learn more about our current offerings here.

---

AI+Edu=Simplified by Lance Eaton is licensed under Attribution-ShareAlike 4.0 International