"...have real-world relevance without importing the outside world directly into every course."
Part 2 of an interview with Priten Soundar-Shah
In the first part of this interview, we began discussing Priten’s new book, Ethical Ed Tech: How Educators Can Lead on AI & Digital Safety in K-12, and some of the insights we can consider for higher education. We continue that conversaton here!
About the Interview series
As part of my work in this space, I seek to highlight the folks I’ve been in conversation with or learning from over the last few years as we navigate teaching and learning in the age of AI.
If you have experiences around AI and education in higher education classrooms that you would like to share, consider being interviewed for this Substack.
Lance: I appreciate your point: how much research is already out there that we don’t act on. It reminds me of a pathway of reading I’ve been doing. I started with Teaching Machines by Audrey Watters that led me to Teachers and Machines by Larry Cuban, then The Teacher and the Machine by Philip Wesley Jackson. And now I’m reading Education and The Cult of Efficiency by Raymond E. Callahan. It was published in 1962 and focuses on the rise of Taylorism in education in the early 1900s.
Seeing this lineage of research and criticism within education, what do you think has held us back from being able to act on it?
Priten: I wish there were easier answers, and I hate being the one to say, “Here are these other ten problems we have to solve first,” but that is the answer here. The lack of compensation for educators does not help incentive structures. That’s a huge component, especially at the higher ed level.
There’s this strange dichotomy where the well-compensated professors are often the ones told their research matters more than their teaching, while many of the people doing the teaching are not compensated enough to treat it as a fully sustainable profession. It becomes transitional work, adjunct work, or part-time work. We do not have structures in place that truly prioritize educators as a profession. We pretend that we do, but I do not think our systems actually do that.
The chances that research and thought coming out of universities, labs, institutions, and think tanks is going to dramatically change the average professor or teacher’s practice across the country—it’s not realistic. We have not built our system to incentivize that at all. There is some complacency there, but I think some of it is warranted.
The other part is that this is a hard project we’ve undertaken as a society. The decision to educate everyone through twelfth grade is hard enough. The decision to also want everyone to get a bachelor’s degree is an even harder thing to take on in a society as diverse and unequal as the United States. That project is difficult, and it requires effort from everyone in the community to acknowledge that difficulty and stop treating it like there are easy answers.
But often the public narrative is, “We just need to do this one thing and it will solve all problems in education.” That’s simply not true. It’s interesting to see some people finally come around to that. I recently saw Sal Khan discussing how he has shifted some of his views.
Of course the technology can do a lot of cool things. This is where I find the disconnect especially high among people focused only on the technology, or people working in tech. The technology is powerful. There is no arguing that. If you want to use it productively, you can.
But the chances that it gets used productively are also very, very low. So there is a gap between can and will. I talk a lot about can versus should, because that’s the important moral question we need to deliberate on. But there is also a gap between can and will.
Lance: Say more about that?
Priten: If a first-year college student wants to use AI to improve their writing skills, it is feasible. The technology can absolutely support that journey. But 90% of students are not going to push themselves harder using the technology. That is not entirely their fault. We have set up incentive structures and assessments that reward shortcuts rather than genuine learning.
Students already wanted shortcuts. We were clearly not getting through to them, and the technology has simply given them one more easy path. I don’t think it has created some massive new crisis that didn’t exist before. We were walking past that crisis and pretending it didn’t exist.
A lot of it was performance. Everyone was going through the motions. One thing I think about a lot is discussion posts and classroom discussions. So much good can happen there. There can be relationality, vulnerability in thinking, curiosity, learning how to deal with disagreement, changing your mind. All of that can happen in a strong discussion forum or classroom conversation.
But we cheapened it until it became: did I post once, did I comment on someone else’s post, did I affirm them, did I speak up twice during class? We incentivized performance, so students performed. That is not on them. We should talk about Harvard University voting on grade caps. I have strong feelings on that. We can bring it in later if you want, but it relates.
Lance: I’m laughing because I took my first online course in summer 2000, and the structure of the discussion was “post once, reply twice,” and that hasn’t changed in the majority of cases. I’m always thinking, the internet has changed—why hasn’t this?
But I want to return to that question of speed and encouraging us to slow down while we have technology that keeps speeding up. I’m thinking about agentic AI and some of the developments there. How might you reconcile that conversation for, say, an institution or a faculty member saying, “We need to start engaging with these tools and getting them into the classroom, because the longer we wait, the further behind we are”? How might you address that?
Priten: There are two different issues here, and that balance has been difficult, especially with people who are new to this and still trying to figure out what the technology can do. I’ve had many conversations with people whose understanding of AI effectively paused in 2023. A lot of educators—especially in K–12, where I’ve spent time recently—explored the tools in 2023, tried them out, decided they weren’t that impressive, identified shortcomings, redesigned some assessments, and then moved on because they were not particularly impressed.
I’ll be in a session and someone will say, “It can’t do math,” or “It can’t look up current events,” and I have to say, okay, wait until I tell you about agentic AI. That has been the moment where I realize things really have moved very, very fast. Because of that, building understanding is important. We should keep up the speed when it comes to helping more educators understand what the technology can do. That is different from how quickly we integrate it into the classroom.
Building tech literacy means understanding what is possible, how society is responding, what the tools are doing, what new harms are emerging, and what conflicts are happening among major companies and institutions. There is a level of awareness everyone should have because these tools affect many parts of society.
But do we need to keep that same speed when it comes to using the tools in our classrooms? I’m not convinced. I don’t think a professor needs to learn about agentic AI and then immediately introduce it into a course. It may change how you think about online assessments or other design choices, but adapting to the existence of the technology does not require embracing the technology directly.
When I say slow down, I mean slow down the assumption that these tools are going to be magic bullets for all of our problems. That’s the part we need to slow down on.
At the same time, recognizing the harms, the ways students might misuse tools, the unique risks being introduced—those require urgency. Strangely, I feel that is the part we are not speeding up. We are not building broad literacy among educators as quickly as we are building institutional buy-in for purchasing tools. There is a major mismatch there.
You’re seeing massive contracts signed in both higher ed and K–12 with tech companies, whether traditional edtech firms or frontier AI companies, but actual usage does not line up. A district or institution signs for a tool, brings in company-led professional development, and then three months later maybe four teachers are actively using it. That suggests we are accelerating adoption without accelerating literacy.
Lance: One thing I’m concerned about this is that students are graduating for whom AI has been part of their education in a more ubiquitous way. Where is speed warranted in that space? We are hearing from students that, as they graduate and interview for jobs, AI is already part of the interview process—more or less, “Tell me about your AI use.”
Priten: This is where higher ed gets a little more difficult, because a good majority of students are headed into at least their first real job at that point. This is also where some honesty about how much of the public narrative is hype becomes important. There is a lot of messaging that says if you want to compete in the world, you need to have learned about AI during your college education. I do not believe that.
My biggest point is that nobody who is currently a leader in the field learned about AI in college. Not a single one of them. Sam Altman did not need to learn about today’s AI in college, and he is doing just fine—however one wants to interpret that.
The skills students will need are not new or revolutionary. They are lifelong learning skills, growth mindset, and adaptability. Those are what matter in a space that moves rapidly. Even if you think about two years ago, the kinds of things you might have told a college senior about using AI effectively or responsibly are already outdated.
Two years ago, human oversight mattered differently. Agentic AI was barely a factor. Context mattered more. Exact prompting mattered more. Much of that is already irrelevant for someone who graduated two years ago if those were the specific technical skills they were taught.
If instead students spent time with deeper questions that the technology raises across their courses, there is a way to have never touched an AI tool and still develop a meaningful understanding of it. We scaffold that kind of preparation in other areas already. We know interviewing matters, for example, but we do not spend an entire college education teaching students how to interview or embedding mock interviews into every class.
Implicitly, though, many things taught in college make students better interviewees. The communicative and social dimensions of schooling help students represent themselves well. That did not require every classroom to become interview training. Things learned in the classroom can have real-world relevance without importing the outside world directly into every course.
This is where slogans become irritating: “real-world education,” “future-ready,” and similar phrases. They pack in too many assumptions. We do not know what the future looks like. We do not know exactly what the real world will demand.
We are better off helping students become resilient, reflective, and metacognitive. Students who understand what they do not know and can keep learning. Those capacities will matter regardless of what the future becomes. I do not think we are doing students a disservice by not wholeheartedly embedding AI skills throughout higher education. I know that is a hot take, and I know there is a strong narrative to the contrary. But if anything, many narrow AI skills may serve only a temporary purpose, while the deeper capacities will be what remains useful in two years.
Lance: There is a lot there that I appreciate about your response as someone with more of a generalist than specialist background—as a history major who somehow found his way here—it makes sense to me. The piece I’m still chewing on is partly about the world we live in. I hear you that much of this is narrative. But I also think there is a fairly intentional effort from industry to increasingly want people to arrive fully formed, without companies investing money to help them bridge into new roles. We all need bridges into new spaces—that is my basic belief.
If abstraction and transfer are true, and if AI can fake so many things, how do people step into those roles with actually demonstrated AI-use skills? That becomes a challenge both for people who have used it and those who have not. If AI becomes a regular part of the internal workplace landscape, how do you speak credibly about a tool you have never used?
Priten: But that is true of most job skills. This is where I’ve been breaking down the things we want to impart to students into three categories: knowledge, disposition, and skills. I find that framework really helpful for grounding any educational work I do, whether it is professional development or our middle school summer camp. The reality is that a lot of the work we do in education is dispositional, and so far we have often treated that as a side effect of education rather than acknowledging that it may actually be central.
When we think about the pace of the skills graduates will need in the economy, many of those can be picked up outside the college classroom. I do not think college has to be the place that explicitly fills that role. You do not need to learn how to use Anthropic’s Claude in a college classroom in order to learn how to use it. That is true of many first-job skill sets. There is a reason internship culture exists, and a reason summer positions exist. People find ways to build the skills they need for the jobs they want, often while learning many other things alongside them.
Learning Claude during an internship is very different from learning it in a college classroom. In some ways, you are probably learning it much better during a summer internship at an investment bank than in a history class. That simply may not be the right setting for it. At some point we have to acknowledge what kinds of things we want college to provide, and what kinds of things we want students to pursue elsewhere. Concrete AI tool skills may be one of those things we are comfortable expecting them to seek elsewhere, just as we have done with many other job skills.
I also think there is a mismatch right now because employers do not really know what to look for. We are in a strange transitional moment. At the moment, employers can get away with a middle ground where someone had a “real” college education, but may also have shortcut their way through the last few years. With the pandemic and AI layered together, no one really knows what a degree fully signifies right now.
As an employer, you also may not know what an entry-level employee will need to do in six months. You may not even need that role in the same way. Everything is moving quickly, and our systems do not move fast enough to adapt every six months or every year. By the time institutions convince people to teach one new thing, they may already need to teach something else, and what was taught in the meantime may already be less relevant. I think that motive eventually crashes into reality.
I don’t think this will remain the dominant narrative in five years. I don’t think we will still be this heavy-handed about whether students are learning everything possible about AI in college. Maybe that is the optimist in me.
Lance: This has been great. I’m curious—any final thoughts?
Priten: The thing I’ve been trying to leave every conversation with is the importance of conversation itself. The most important thing all of us can do right now is talk to each other more, spark more dialogue, and admit what we do not know. The pressure to feel like you have all the right answers in this space is really high, and I think the tech ethos makes that even worse.
As educators, we usually have more comfort acknowledging what we do not know. That often comes with the territory—especially in academia, though maybe not always. But some level of humility is more normal in education than it is in tech. Right now, though, there is urgency to appear as though we have all the right answers so we can navigate a very uncertain period.
The more honest we are that we do not have all the answers—that this will be a rough journey, and that there is no magic fix coming in September that suddenly solves education—the better off we will be. Some camaraderie in that uncertainty will help people think more freely. Right now, if you feel hesitation, you often keep it to yourself. Then it stays there. You do not get to explore it, sit with it, or think through real solutions because you are a little afraid or ashamed to admit it publicly. On the other hand, if we talk with each other, we might realize that we do not have to get through this alone, but in fact, with an entire community of like-motivated educators.
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)
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
 | A guest post by
|