A growing number of colleges are hiring Chief AI Officers to try to help set a strategy for adapting to generative AI. These ‘AI czars,’ as they are often called, are in a unique position to hear the good, the bad, and the messy about what's happening with AI in education. For this episode, Jeff talked with two AI czars — both from large state universities — to hear what they’ve learned. And we talk about why both of them have left the role.
This episode dives into the experiences of Chief AI officers at universities, including some misadventures and lessons learned.
“I’m an ‘AI Czar.’ This is the Role,” by Jeffrey Bardzell.
"AI-Enhanced Pedagogies: Rethinking Learning, Curriculum, and Human Potential in the Age of Intelligent Machines,” by Alexander (Sasha) Sidorkin.
"The Rise — and Fall? — of the AI Czar," in The Chronicle of Higher Education.
Jeff's new Jagged Intelligence newsletter about AI on campuses.
Thanks to this episode's sponsor, Studiosity.
This is an AI-generated transcript so there may be some errors.
Sponsor:
This episode is brought to you by Studiosity.
Real learning happens through real thinking and real experiences. It's something that takes effort, not shortcuts. For over 20 years, Studiosity has partnered with educators to protect the thinking, the essential work that the student must do for themselves, not detection, not an us versus them culture in teaching and learning, just the real journey behind every degree that is worth holding. They're doing it for the institutions that trust their students, for the institutions that back their teachers to teach, not police, and for the institutions that stand behind their degrees keep the learning real with Studiosity.
Get the walk through at studiosity.com/learningcurve.
Jeff Young:
Hello, and welcome to Learning Curve, where we look at how education is adapting to the world of generative AI.
I'm Jeff Young, a freelance education journalist.
Colleges are starting to hire AI czars. This new job aims to coordinate all the AI activity going on around campus, which makes it a pretty challenging gig, since there's still no agreement on what colleges ought to be doing with AI, or whether it should even be involved in education.
The official name for these jobs are Chief AI Officer, that's what's on the business card. If people still have those, but informally they are known as AI Czars. Ever since I first heard about these jobs, I have been curious to talk to the people in these roles about what they're learning. After all, AI czars at colleges are in a unique position to hear the good, the bad, and the messy about what's happening with AI in education.
So, for this episode of Learning Curve, I connected with two college AI czars who have been vocal about sharing the lessons that they've learned, actually, both of the people that I talked to, they are no longer chief AI officers. They recently finished up their jobs, and the fact that they've moved on, it makes it easier for them to talk more openly about the challenges they faced.
To start off, I got on a Zoom call with Jeffrey Bardzell, who was the first Chief AI Officer for the University of North Carolina at Chapel Hill. He's moving on to another university at the end of May, and he's been writing about what he's learned about being a Chief AI Officer on his blog called Interaction Culture. The title of his post about all this is called I'm an AI Czar. This is the role. I'm going to link to that post in the show notes, and I absolutely recommend it. It is meaty and packed with insights, and as soon as I saw it, I really wanted to get him on the show and talk to him more.
I started by asking Jeff Bardzell what he sees as the main problem that chief AI officers are trying to solve.
Jeff Bardzell:
I think the single biggest challenge is that you've got three three groups of stakeholders that are really in a very different place, so you have you have administrators like presidents, chancellors, and provosts, and they are under incredible pressure to do something to be relevant to prove that, that, that they're providing ROI for the public, or for the citizens, or for students, or, and so on, and so they're like, we have to do something, we have to be relevant, this has to be yesterday, catch up, catch up, you know, and they're, they're in that position, and then you, you have the, the faculty who, who are seeing what AI is actually doing to teaching and learning, and, and, and I don't think this is the final position on, on this, and I, and I have lots to say about this, but where many faculty are right now is they, they are seeing how AI is being used to circumvent the whole point of learning and to circumvent the whole point of pedagogies that have worked for generations, and they're they're harmful to students and they're harmful to the disciplines that they're trying to advance, and, and again, I'm not saying that's sufficient to sort of sit down and say we refuse to deal with it, or let's just ban AI. Don't at all support that view, but I do want to recognize that those are pretty important points, and they have to be taken seriously. And then you have students who are sort of caught in a position where, on the one hand, they're told entry-level positions are being wiped out because of AI. So, students are really struggling to find jobs, and you had better know AI, and you'd better be good with AI. AI is the future. Catch up, catch up, catch up. They hear that message, and then they also hear, but if you use AI here, you're going to find yourself on the wrong side. An academic misconduct hearing, because you're a cheater, and so students are navigating a pretty fundamental contradiction, and that's not acceptable or sustainable, like we have to do better with that, but for me it really comes down to the faculty, they are the most important figures in this, and they are right to be to feel disrupted. They are right to feel that this is existential, but how to actually support them in moving forward? I think that has been a challenge, and I think I had some ideas, and I tried to communicate those on the on the blog.
Jeff Young:
One of the things that I was interested in, in the way you framed it, is that you talk about AI changing what it means to have expertise. So, can you say just a little bit more about that, because that seems like key to even, you know, what can be done here?
Jeff Bardzell:
I think that's at the center of my argument, actually, and I think so. Let me start here.
I think one of the misconceptions that people are working with is they're thinking of AI predominantly as a technology and predominantly as a topic in its own right, and analogously, if you think back to the 1990s there used to be people who are like, "I need to take a course on using computers. I don't, I don't know computers well enough, like community colleges were offering them, and I, you know, I remember talking to people, and my mother would say things like, "I need to learn computers, and today that sounds quaint, and I think we're doing almost the same thing with AI today, we're like, I need to learn AI, you need to learn it, students need to learn AI.
And I actually, there is some sort of table stakes level where that's true, but above the table stakes stakes level, I think that becomes less and less productive, and so what I would, I try to argue instead is that AI is enmeshed in the in the the topics that we of inquiry in the methods that we use to pursue those topics in the way that we interact with the data that we collect with those methods, in the way that we analyze that data, in the way that we produce research products, the way that we disseminate research products, and also the ethics by which all of all of these activities are are overseen and ensured that they're they're rigorous and fair and just, and it's that enmeshment I think that that the dis that faculty and the disciplines are by rights the true experts and and I think that's what that's what constitutes expertise. What constitutes expertise is is being inculcated into a sophisticated practice, understanding its ideas, understanding its methods, or its behaviors, understanding its ethics, and that's what comes to constitute expertise, and what I think is that with AI, the nature of that expertise is going to change, and this is not about displacing humans or any of these other kinds of things. This is just more about shifts in how work gets done, shifts in what we are able to automate and what we choose not to automate, how we make those kinds of decisions, and how we train people to do the work that we don't want to automate, and how we train systems to do the work that we do want to automate, and so I think, like, that's the work that that we all need to be doing, I think, and I think industry is doing it, and I think higher education has a role to do it specifically in the context of advancing disciplines, so you know, I'll let, for example, UX designers, let them figure out AI in the context of UX design, but in the context of the formulation and theorization of human computer interaction as an academic field, that should be done by faculty in human computer interaction by the international research community that identifies with that and has been trained into that, if all that makes sense.
Jeff Young:
Yeah, what I, what I kind of took from what you're saying, and then what you wrote is that there's a certain historical mode that professors should tap into here around AI, not to just be somehow reading every new article that comes out about AI, so they can understand what you know, the latest version of Claude does that it didn't do yesterday, but to be in touch with their own professional world or discipline and just. Like they've had to do with everything, not just technology or AI, but just how are people in the practice of doing whatever it is they do, whether they're a physicist or a marketing professor, or just like everything that professors might do, whatever discipline they're in, what is their discipline doing with AI, starting with that as a way to make sure they're up on it, just like they were up on things that aren't as controversial in some cases as AI or new.
Jeff Bardzell:
Exactly, I do think it's not a fair ask to expect faculty members outside of computational and technical disciplines to become amateur computer scientists, and I think some of them are hearing that message. It is a fair ask to say, "Oh, you're a historian or you're, you're, you're a sociologist. What are people in history, faculty members in history? What are the historians talking about with regard to AI, and then how should AI interact with your field?
So it's not just a reactive question. Oh my god, we're being hit by AI. It's a proactive question. How do we want AI to interact with, or contribute to, or help us advance this discipline? And equally important, How do we not want that to happen, with the caveat that you're not allowed to say we don't want it to happen.
And then so I think the field should be asking those questions, and then I think each individual faculty member should have their own principled take, like I don't always agree with my field, and so I think it's fine if a historian doesn't agree with the general direction of their field, but whatever their principle, they should know what their field is saying.
That's one.
Two, they should have their own response to that, and three, whatever their response is should be reflected in their research and teaching. And so I think that's a fair request. I don't think it's a fair request to say you need to learn AI, which actually is almost meaningless, because there's no consensus on what that even might mean, and I think people think it means a lot more than it really does.
Jeff Young:
Wait, you don't know what it means to learn AI?
Jeff Bardzell:
People in the C suite, anthropic and Open AI…
Jeff Young:
And that's because it changes so fast, and what, what are the reasons, or is..
Jeff Bardzell:
Because I think that's the wrong question.
It's like asking, what does computing mean? What is it? What is it? What is sufficient computing? Now it's just, it's just the wrong question. An air traffic controller needs to interact with computers in, you know, in a very different way than an accountant does, and so it's to me, it's not really even useful to talk about them both doing computing, because one of them is really doing accounting, and that's the heart of it. And then, how is computing or AI enmeshed in, or supporting that activity of that practice? And the same thing with air traffic controllers. So, I think we've got it backwards. I think we keep talking about AI, and we keep talking about AI fluency, but what we really need to be talking about is what are the practices that AI is enmeshed in and is supposed to be supporting, and how do we ensure that those practices continue to be done in a rigorous, high-quality way, and also, How might we advance them so that they're done in innovative and exciting new ways, and help us move forward. And I would much rather see faculty engage in those two questions. And also say, when I go to faculty, they start - many of them start in a defensive crouch, expecting me to say, "You're behind, you need to catch up. This is the future, and instead I say these other things. I say, like, forget about that. Think about your discipline, and think about how AI is already entangled in it, and how you think it ought to be. And when you do that, they start, they change, it's disarming. And then some of them even start to get, get pretty curious about it, and like that's that's what we need to be doing.
Jeff Young:
There's… I'm curious, you know, I always love stories and anecdotes. I mean, you have this intense, you know, intelligence gathering you've done, and knowledge, and talking to so many people at this, you know, big state university. I'm wondering, I know you put a couple of these kind of vignettes, or like little hints of like moments that were surprising about AI, or how it's touching so many different aspects of the job of a faculty or administrator. Could you go over a couple of the surprising things you kind of learned or saw about AI use that might go beyond, you know, yes, a student might have used it for cheating, and it's hard to tell. I mean, that that is a known issue out there, but what are some other ones you, you heard or encountered?
Jeff Bardzell:
I talked to a faculty colleague who had to write a paper that was not, it was not central to their we. House, and it was not like in a top-tier journal, so it was sort of a something that would be useful to do, but not a little bit on the side, and they decided, as an experiment, to really throw themselves at a new compositional process, as just experimentally to see where they landed, kind of in the lower stake situation
Jeff Young:
With AI?
Jeff Bardzell:
With AI, and so they apparently they, they just talked, they talked about what they wanted to talk about, they talked about their ideas, and recorded that, created a transcript out of that asked AI to clean up the transcript, and then they then printed that out, read that back to the AI, but curated and edited as they were reading it, and then so they had a newer, a second iteration, and kind of in this way the sort of dialog strategy they went back and forth, just giving content, letting the AI capture it, clean it up a little bit later, starting to ask questions about how I want to make this argument, how do I make this the best way? Should I be moving this over here, moving that, and letting AI help that? But it was very much the point is, it was super conversational. They never, at no point, did they cede authorship to AI. They just used AI like almost like a grad student, almost like an assistant. Can you be another set of eyes? Can you give me some pushback on this? And I, you know, I had a similar approach. I had to do a presentation to the Chancellor last December, and I made, I made all my slides, and I thought it was in good shape, and just on a lark, I uploaded it to, I think, it was perplexity, and I said, I've got to make a presentation to the Chancellor. Can you evaluate this on those terms?
And it came back, and it was like, whoa, you got some work to do, my friend. And, and I read through, and I was like, actually, I agree with 80% of what it's suggesting, like I actually could have gotten to the point faster, I could have been more direct, I could have exemplified this better. And then I redid the whole thing, and then I resubmitted it. And the second time the advice was 20% useful, and I thought, aha, that's how I know I'm getting better, because the first time it had lots of really useful practical advice for me, but the second time it had real diminishing returns, which meant that it still had to say stuff, but it couldn't come up with as much useful stuff, because I infer there was less useful stuff to come up with, so these are, I think, you know, these kind of interactions where you're not seeding your authorial, your authorial voice or intentions to AI, but you're letting AI be a practice reader for you, be a test environment. I think those are that's an example of a benevolent use.
Another example I heard was that students are uploading their lecture notes, and they're creating mock exams with AI, and then they're taking those exams, so that when they get to the real exam, they really have some confidence about how well to do, and it works. And you know, when I talk to faculty about it, they're like, yeah, that's not cheating. Most of them don't think that's cheating. And so I think a big piece of this is just, can we can we disseminate benevolent use cases that people feel would feel comfortable with, or that more people would feel more comfortable with, so that the narrative is not always a student used an LLM to write an essay, and just skipped learning. I mean, certainly that use case is there too, but there's.. there are benevolent ones, and I think it's really important to get them disseminated, so people can understand.
Jeff Young:
And I guess you also named some things where there's not a way to avoid this. It seems like one of the messages from your pieces that you can't just skip this technology as a faculty or a student, I guess, but the faculty, because I guess there are some things that faculty were encountering that were challenges, whether they expected it or not, on things like publishing, even some scholarly publishing issues.
Jeff Bardzell:
That's right, so there's there's a lot of rhetoric of AI about what's going to happen, so it's in the future tense, but the reality is it's empirical, it's already happened, and so yeah, so for example, a lot of academic journals are seeing huge upticks in submission. Is, and I think we can all guess why that's happening, and some of it may be nefarious, and some of it may be benevolent.
Jeff Young:
It could be like that colleague of yours that was, you know, having it help him coach him to write an essay he might not have written exactly.
Jeff Bardzell:
Or just write it in two weeks rather than two months, or something like that. But every time you have a submission, if you've got to send it out to three peer reviewers. If you get three times as many submissions, that's nine times as many peer reviewers. And so finding reviewers for papers is becoming a crisis. And the practical response that reviewers are put in is they start to have to use AI to help them do peer reviewing, and then that freaks people out, understandably.
Jeff Young:
Yeah, because people's careers are on the line with publishing,
Jeff Bardzell:
Not only that's true, but also the integrity of the discipline itself is on the line. It's real, it's existential, so you know. So this is why I'm like, this is why you can't just ban it or have these overreactive policies that, that, that, that say you're no one's allowed to use AI.
The fact is, you got auto complete in the world. What do you think is driving that? You've got people using Grammarly. What do you think that’s like? Where you draw the line, people are using AI all the time.
So it's the real question again, is how do we use it well in the context of maintaining the integrity of disciplines and advancing those disciplines, and I don't, I'm just not hearing that question being asked enough, because the question is always, how do we catch up and make sure that we're covering AI, and I just think that we've created a red herring for ourselves that's demoralizing, disempowering, and barking up the wrong tree anyway.
And so that's why I keep saying it's like it's about the disciplines, it's about expertise, so it's like James Carville, it's the economy stupid. I wanted to write a blog post at one point called "It's the Expertise Stupid, but I thought that might not go over well.
Jeff Young:
Yeah, I thought for a second about calling the whole episode AI czars say it's the expertise stupid, but I talked myself out of it, Bartzel is leaving the university at the end of the month. He'll be moving to Worcester Polytechnic Institute, where he's going to serve as the dean of the school of arts and sciences.
Jeff Bardzell:
You know, the reasons for the move, as you might guess, are complex, and some of them are personal and local. So, I'm just going to try to generalize to the points that might be relevant to your, your, your audience, people talk about AI readiness at the organizational level. You know, I was reflecting on this this morning because I was anticipating this question.
I was thinking there's also maybe AI czar readiness within an organization, and I think I think a lot of universities are trying out this position, but I think because the position is new, a lot of, a lot of universities' organizations don't quite understand what they want it to be, and and there's a lot of working parts, and so I, you know, I ended up in, in, in this role where I was told before I started, for the first three months, a big part of your job is to figure out what the job is supposed to be, and that sounded creative and fun and appropriate, like what I should be doing, and all that, I stand by, but in hindsight.
I also recognize that sort of tacit in that statement is you're also going to kind of start from zero, you're not going to have any real authority, you're not going to have any real budget, you're not going to have any real resources, because nobody yet knows what the position is, and so some of that work to figure that out turns out to be sort of political and organizational and not super intellectual.
And I'm somebody who, I love to be in the weeds of research and the weeds of curricula, I love the content of what people are working on, and I, and I found myself in a position where I was like navigating for resources, and and I think one of the practical challenges is when you create a senior position out of nowhere, there are other people who had been doing work that all of a sudden now is possibly in your, in your remit, and that creates some organizational challenges and negotiations as people figure out, people figure out who should be doing what, and so there were some challenges there, because I don't think that at the senior level of the university a lot of people had fully appreciated the implications, and I think people thought they were on the same page, but when once. Got into the weeds, they weren't on the same page, and so that created some practical challenges that had to be navigated
Jeff Young:
After the break, lessons from another AI czar, one who faced objections from the faculty union over an AI experiment that he tried.
Stay with us.
Sponsor:
This episode was brought to you by Studiosity. Real learning needs a struggle, but for educators, assessment security just means more detective work, and that's not what students came for. It's not what the teachers came for. True learning and authorship isn't found in a probability score or a surveillance log, it's found in the journey. Studiosity makes assessment security easy by shifting the burden away from every educator and into a proactive institution-wide infrastructure, from Viva style inputs to clear binary validity reporting. Studiosity is the multi-layered assessment security process, turning institutional policy into practice. Just a clear signal to the teacher, owned by the student, that I did this thinking. De-risk your assessment flow and reduce workload. Give your teachers back the time to teach, not police, and give your students the agency to show their own growth. Keep the learning real with studios. Get the walkthrough at studiosity.com/learning Curve. That's studiosity.com/learning Curve.
Jeff Young:
I’m excited to announce a new project I’m doing that I think you’ll find useful. It’s a new newsletter about AI on campuses that I’m writing for The Chronicle of Higher Education. It’s called Jagged Intelligence, named for a term that captures the uneven quality of generative AI, since it’s capable of such amazing feats and spectacular fails, and we’re all just trying to figure out how to adapt. For the newsletter I offer original reporting and analysis looking at emerging best practices and cautionary tales to help you understand what this technology could mean for the future of higher ed. Just go to Chronicle.com and click on ‘newsletters’ at the top of the page to sign up.
The other AI czar that I wanted to talk to you was one of the very first in the country to hold this job. He's on the other side of the country at Sacramento State University. His name is Sasha Sidorkin, and he's now a professor of graduate and professional studies in education at the university.
He became the university's first AI czar back in 2023 which was pretty soon after ChatGPT was released. I first met Sasha Sidorkin last year when I wrote a big report for the Chronicle of Higher Education called “Leading in the AI Era.”
I was struck by his story, so I wanted to talk to him again for this podcast.
He believes higher education needs to make some big changes to meet this moment, and he outlines them in his new book, “AI Enhanced Pedagogies: Rethinking Learning, Curriculum, and Human Potential in the Age of Intelligent Machines.” One of the experiments that Sasha Sodorkin tried when he was AI czar involved building custom AI chat bots for professors that were trained only on the materials from their courses to offer as a resource for their students.
Sasha Sidorkin:
Yeah, it became sort of very early clear to me and to others as well as that one of the sort of the key applications of this new technology would be a classroom assistance or custom bots, I tested that extensively, and I also taught a course I tested in my own course as well. So the technology is really well suited for this function, where it can give additional help to students and where it could both explain the syllabus, but also answer questions about the content of the course, and also help them with their homework.
Jeff Young:
But when he sent out a call to faculty, he was surprised by the reaction.
Sasha Sidorkin:
Then I kind of sent out a blanket, you know, offer it to all faculty, saying, well, if you want to build ones, here's what I need from you. I need some materials, I need your syllabus, I'll build it, and then I'll give you the link, and then you, you know, make it available to students. So, and if I remember correctly, maybe about 18 or so people volunteer to do it right away, like within the first day or so of this whole thing, and then then what happened is that the faculty union, as they say, registered concern with the administration, because they thought it was replacing faculty labor, so and the administration asked me to take it down the whole thing, so and I had to actually kill already functioning bots and just close them down because, because of that concern, now people are still free, were still free to use their own what they built, but they did not allow me to do that for other, for other faculties as well.
Jeff Young:
When I wrote that AI report for the Chronicle of Higher Education, I talked with one of those union representatives, Patrick Oberly, an associate professor of geography at the university. He told me that the whole effort felt like asking professors to train robots to replace them. He said, if a faculty member designs a chatbot with their own materials, what's to stop the administration from. Not teaching their course in the future, or saying we don't need a faculty member, we'll use a chatbot and a textbook and an exam. I ran that concern by Sasha Sidorkin to get his response.
Sasha Sidorkin:
Well, no, I think the concern is really, really misplaced, and it's based on really misunderstanding what technology can or cannot do.
Jeff Young:
Okay,
Sasha Sidorkin:
and unfortunately I have to say that's not just the unions, every every critic that I hear that are adamantly against use of AI, from the context of the critique, I always understand you don't have much of a personal experience dealing with the technology, otherwise you wouldn't have said this, right?
Jeff Young:
So walk us through why that is. How do you mean?
Sasha Sidorkin:
Well, like for example, people say, 'Oh, well, all content will be AI-generated. Well, you don't understand how that works, because the value is created on the interaction of human with the machine. There is no value in AI alone. It just doesn't produce anything of worth if you don't actually input your ideas structure if you provide intent, if you, if you don't provide evaluation of the output, right, and you can only imagine that if you may be experimented with it once or twice, and we're awed by how wonderful you know it is.
So, it's very superficial knowledge of the technology that leads to those concerns, and another, I think this misconception is that people like the union leadership, they don't understand that education is not an informational industry, it's a relational industry. Students come to us because they want to enter into relationship with somebody, with the grown-ups, with each other, so they're there for the community, not just for the, you know, you can get information anywhere for years now, for decades now. You could, if you really were an other didact, you could teach yourself anything you want for the last 30 years or more.
So, and I think they kind of sell their profession cheap when they think, oh, the bot can replace me, so it's really not the case. And, in fact, I think that I'm a union member now, and I think what we should be doing, we should be negotiating for more support, more training to give us some time and space to learn about how we can use AI in our teaching, rather than cutting us off from the technology and making the profession less relevant, appearing to be more complaining to the public, and actually undermining further the trust in higher education that it's already been suffering in the last few years.
So, yeah, I think it's an unfortunate strategy that has very little merit. I don't see any foreseeable future that would replace teachers with AI, it's just not going to work. The technology is not, it's very, very far from there.
Even if you could do it, then the students would bail out. They don't want to be taught by a machine, they want to be taught by a human being. But if the machine will assist them and make them more effective and more efficient, that's great, you know, it's question is about can we amplify our abilities with AI.
Jeff Young:
I see both sides of this one. On one hand, professors worry that some administrators might be willing to try to replace them with AI, even if that professor doesn't believe that the education would be effective that way, and folks like Sasha argue that AI can actually enhance the human side of learning by doing things like providing tools for students to look up basic information on the course. The challenge for any college AI czar is building enough trust to hold discussions like this about the pros and the cons of any AI experiment. So, what about the idea that Jeffrey Bard Zell had about really focusing on disciplines in driving discussion about AI.
Sasha Sidorkin:
Yeah, absolutely. If there was a generic solution, he and I would be rich by now, because we would have sold it to everybody else, and they would have a.. no, I mean, the kind of a curriculum restructuring that's a English composition, college composition courses have is very different than what computer science is dealing with, right. So you cannot have like a generic solution. I wish the professional organizations nationwide would invest some of the money into figuring out, because you know all math instructors in the nation could probably get together and try to figure out the new curricular sequences, but you know, I have to admit that they don't show much leadership. They will have some task force, you know, trying to try to regulate, like, write little regulations or policies, which I think is really not a useful thing to do instead of revising curriculum, which is what they should be doing.
Jeff Young:
Interesting. Well, I mean, I guess you mentioned money too. I mean, these are not necessarily like rich organizations.
Sasha Sidorkin:
But still a cooperation across different campuses would probably be economically more feasible. They could I. Have you know a task force that would just review a typical math curriculum or typical geography curriculum or history curriculum and say what's the new sequence, what news? The key point is here that the learning outcomes are changing, that's that's the most difficult part.
Jeff Young:
How do you mean?
Sasha Sidorkin:
Well, what you need to know and be able to do in every discipline is changing, right?
Jeff Young:
Because the job market is changing, in the way research is working is changing.
Sasha Sidorkin:
Yeah, because people start using AI in their work, so, like, say, computer science, like computer science, probably the most typical example, because they're like also directly affected, just like us, so in computer science, like basic coding is probably need to take a different shape. Nobody codes by hand anymore. They all use agents, and they all kind of use, but the skills to supervise agents to give them, those are new skills that they should be teaching, right? So they need to de-emphasize the procedural skills, like basic coding, and reemphasize, like huge, hugely reemphasized the new skills, right? And that calls for once you figured out your new learning outcomes, then you need to kind of deconstruct your curriculum and reconstruct it again based on this new outcomes, but it's not just computer science, and every discipline is like that, you know. People say in teacher training, teachers are not writing lesson plans by hand anymore, right? So they are the, but we're still teaching them how to write lesson plans. Like, why are we doing that?
Jeff Young:
You mean they use AI?
Sasha Sidorkin:
Yeah, they should go to magic school, which is like in bulk of teachers are using and learn how to operate with it, learn how to adapt it, so that it writes lesson plans relevant for your particular class, your kids, how to fit in formative assessment data, and all that. Those are different skills. There's, there's a different skill than just, you know, devising the unit, looking at the standards, and then creating a lesson plan, so that's what really the bulk of disruption is coming from, because we can't teach the same things anymore, we have to teach something else.
Jeff Young:
Sasha's position as AI czar was eliminated last year for budget reasons, so they don't have any eyes are anymore at the university, and now he is back as a faculty member. I asked him what he took away as the biggest lesson from his experience.
Sasha Sidorkin:
What I think I've learned is that ambition does not equal a good plan. So, yeah, you have all the best desires, but if you don't have resources to support it, you can't do it, and of course, the most, I think, university presidents across the.. I mean, I don't have data for that, but my guest, I mean, I give a lot of talks, most of them do not realize the scope of the crisis that they were dealing with, because they have other big crises, you know, they have everybody has budgetary pressure, everybody has political pressure from federal government, and many cases state governments as well. So they're so preoccupied with that, that they don't really see the big picture. Is that if we do nothing about AI, we lose more credibility with the public. I mean, that's the, that's the main risk, because you will have degrees that are diluted, that's easy to do, so more universities will kind of collapse into diploma mill situation, where you know, you give essays and a ChatGPT writes it, and then another ChatGPT will grade it, and that's all done, so there are there are some bad scenarios in this whole thing.
Jeff Young:
Yeah, so you're saying the stakes are high for paying attention to this.
Sasha Sidorkin:
Well, the stakes are high, and they're like gradual. It's not going to happen overnight, but if we lose face, if we lose credibility, like we can't cope with that, public will ask, why should we support the higher ed if they can't even figure out that, so yeah, there is definitely a risk, and it's an additional factor. It's not the main factor.
I think there's a lot of doubts about value of higher education and general economic value of it. There's also some parts of the population that are unhappy with our ideological bias, so this was another extra one. It's another one. If we screw it up, we're gonna lose, not maybe the main one, but still a significant factor.
Jeff Young:
Yeah. Do you have any advice for other AI czars at other universities?
Sasha Sidorkin:
Well, I think my advice would be to go after the core problem, which is the curriculum, that's the, that's really the biggest one. So, they need to encourage faculty to revise it, because plugging in AI into the existing curriculum is really damaging curriculum, it doesn't do any good. So that's, that's, I think, is the main case. You can kind of superficially adopt some AI elements, but if you don't revise your idea of the goals of learning or learning outcomes, it's all going to be temporary and not significant. So, yeah, go, go to the hardcore problem there.
Jeff Young:
So, it's not just putting a little assignment to help people understand how AI works,
Sasha Sidorkin:
Yeah. No, that's not gonna work here. Yeah, AI, yeah. Well, I mean, the AI literacy is actually easy to teach, like engineering terms. You can do it. People, I've seen the courses, I tell the course myself, like that. Sure, it's really not that hard. It's really not that hard. You just so when it's add on, but when it conflicts your rest of your program with the university, where people say don't use it, then it's really a credibility problem.
Jeff Young:
Both of these former AI czars argue that colleges need a plan, they need a strategy to adapt to how AI is changing knowledge work, but their stories also illustrate how truly hard this is right now.
This has been Learning Curve.
We are trying to sort through what happens to education once AI sweeps in.
Please follow Learning Curve wherever you listen, and as this academic year wraps up, I am asking you to grade me well. Give me a review, a rating on your favorite podcast app, please. This episode was put together by me, Jeff Young.
Our theme music is by Komiku, with additional music by Blue Dot Sessions.
Art for the episode was generated with Mid Journey.
I also want to thank the editors at the Chronicle of Higher Education. It's been a treat to be back working with them on my new AI newsletter, Jagged Intelligence, and I will be sharing some of what I learned there for this podcast, of course, and vice versa.
Thanks also to the sponsor for this episode, Studiosity. We'll be back in two weeks.
Until then, thanks for listening.