Mushtaq Gunja: Hello and welcome to dotEdu, the public policy podcast from the American Council on Education. I'm Mushtaq Gunja, and I'm here with my co-host, Sarah Spreitzer. Sarah, how are you?
Sarah Spreitzer: Good, Mushtaq, so happy to be on this Jon-free intro to the podcast.
Mushtaq Gunja: Where is our illustrious co-host, Jon Fansmith?
Sarah Spreitzer: He is taking a well-deserved vacation because I thought he probably needed a break before we find out what happens in the midterms.
Mushtaq Gunja: Jon and I, by the magic of having had this conversation a little bit earlier, talked to Paul LeBlanc, former president of Southern New Hampshire University, author of a new book on AI and about how higher education is going to work in a world that's transformed by artificial intelligence, super intelligence. And we're going to bring that conversation to you in just a second, but we are recording this on Wednesday, October 28th, right before the midterms. And Sarah, they're coming next week. How are you feeling?
Sarah Spreitzer: Oh my God, I just want them to happen now. And I know next week we're actually going to be talking about the midterms right after they happen. But DC, it feels almost like it's in a panic. Every day I read new polling data coming out. But Mushtaq, I know that you follow all of that stuff super closely. On Tuesday, I will probably be in bed watching the election returns and will be expecting to text you, but what state are you watching most closely in the lead up to the midterm? And then on Tuesday, which one will you be paying attention to?
Mushtaq Gunja: Ohio has two interesting races, both the governor's race and the Senate race. And on the Senate side, it's one of those states that's really near the tipping point for whether the Democrats will control the Senate or the Republicans will. So the Sherrod Brown, Jon Husted race, very, very interesting. And then of course up at the governor at the top of the ticket, we got Amy Acton and Vivek Ramaswamy. And somebody stopped me on the street the other day and thought that I was Vivek, and-
Sarah Spreitzer: No, no. Are you serious? You don't look anything alike. [inaudible 00:02:38].
Mushtaq Gunja: We don't look anything alike. And my hair is much less tall than Vivek's is, I think.
Sarah Spreitzer: Yeah, his is really poofy. He has the pompadour.
Mushtaq Gunja: But Ohio, it's on the East Coast, or at least on Eastern Time. And so we'll get those results a little bit quicker. We're going to have to wait for quite a bit of time on Alaska, both because of time zones, but also because of ranked choice voting. California's going to come in and the ways that California comes in. So that's going to be a while. But these East Coast races, I think could be really telling for us. So Ohio is definitely one I'm going to be watching very, very closely.
Sarah Spreitzer: Yeah, I think Michigan too. I'm really interested to see where Michigan falls, because I think that that one's gotten a lot of national press.
Mushtaq Gunja: Yeah, absolutely. But the Michigan race and a couple of others illustrate, I think a little bit of the sui generis version of what is happening writ large. So let me just say, the way that I think about these numbers are in 2024, Donald Trump won the popular vote by a little bit more than a point. The Republicans won the congressional vote to the extent they're a national congressional vote. They won it by about 2.6 points. So that 2.6 number is actually a really key one, because maybe the number I'm watching closest, Sarah, I don't spend as much time looking at any individual matchup poll. In the Sayed-Rogers matchup, I mean, you can look at the average of polls, but actually the thing I'm looking at more is a little bit of the generic ballot number on... So most pollsters will ask, "Are you planning to vote for the Democrat in your House race or the Republican?" That number right now in an average of polls, and I think in an average of something like 14 polls over this month, is Democrat plus eight.
Sarah Spreitzer: Wow.
Mushtaq Gunja: So Sarah, what that means is that we're going from an environment, if the polls to be believed, from R plus 2.8 to Democrat by eight. And so that's a shift of about 11 points. That's the heuristic that I've got in my mind when I'm evaluating all these states, we should expect some amount of shift. Now, if it were a uniform shift, we'd see 11 points across the board. But what I think we're seeing is candidates mattering quite a bit. So in some states that shift appears to be something less than 11 and in other states it appears to be a little bit more than 11. So take Maine for instance, this is a state that Kamala Harris won by something like six points. You could theoretically add that 11-points shift to the six points, and you might expect a Democrat to win that race by 17 points. Troy Jackson is not, I feel pretty confident, is not going to beat Susan Collins by 17 points, right?
Sarah Spreitzer: Mm-hmm.
Mushtaq Gunja: Because of something around the environment in Maine and something around candidate quality, both on the Collins side and probably on the Troy Jackson side. But when we're dealing with a generic R or generic Dem, and in the House side there's a little bit of that, I think we should expect to see an 11 point race. Now, Sarah, you asked about the House. Let me just note one thing. In 2024, there were 15 races where the Republican won by five points or fewer. There were 21 races in which the Republican won by eight points or fewer. There were 23 races in which the Republican won by 10 points or fewer.
Sarah Spreitzer: Okay.
Mushtaq Gunja:
Okay. So if you're thinking about a 10-point shift leftward, you might expect, if that shift is uniform across the 435 districts, that Democrats might at a baseline be expected to pick up 23-ish seats, something like that. Now you got to add in gerrymandering, complicated. I mean, the Republicans made some significant moves on gerrymandering. So that number might come down a little bit from 23 to 18, 17. There's five, six, seven, eight seats that I think probably the Republicans are going to gain because of some gerrymandering. But that's about what I'm thinking about from a House perspective. Now, if there's a big polling error, we shouldn't expect that that Democrat plus eight is going to be precisely right. I mean, it's over a lot of polls, but it's probably not going to be precisely right. The question is just how big an error might we see and in which direction might we see it.
Sarah Spreitzer: I'm just going to go with my gut, Mushtaq. I'm just not even going to listen to any of the polling data. But given your crunching of the numbers, are you ready to make a prediction, which I will hold you to next week, on the House and the Senate, which will be the controlling party and how many seats do you think the Dems and Republicans will hold? Are you convinced that the House is going to flip and do you think there's a chance in the Senate?
Mushtaq Gunja: Do you think the House is going to flip? Just as a baseline, let me note that in 2018, the first Trump midterm, after the midterms, the Democrats controlled 235 seats, I believe. And the election that I think this most resembles is the '06 election, Sarah, that you remember so well.
Sarah Spreitzer: Oh, so well.
Mushtaq Gunja: The midterms in the second term of the Bush administration, you might recall, unpopular war in the Middle East. You might recall a whole bunch of corruption scandals, especially the affecting House members. That's a little bit different, most of the corruption's around the executive here, but that was a race that was dominated by the war and corruption. And the Democrats did very well. They ended that election with, I believe, 233 seats. So that's what I'm looking at. I would guess that the Democrats take back the House with about 230 seats. So a margin of-
Sarah Spreitzer: Not a big majority?
Mushtaq Gunja: Well, I guess it depends. We've had such narrow majorities the last two, three cycles, the 220s, 222s. I mean, 230 gives a decent amount of breathing room, but we'll see. That's about what I would expect. On the Senate side, a lot of races that appear like they're pretty close. Look, I think that the Democrats are almost certainly going to take the North Carolina seat. They need four to get to 51. They're almost certainly going to take the North Carolina seat. And I would say there are eight to nine races that are pretty true toss-ups.
I think in a world in which they are pretty true toss-ups, I might think Brown is slightly favored in Ohio. I might think that Ricketts is slightly favored in Nebraska, but I would say maybe there are nine-ish seats that are toss-ups. You might as well just give half to Democrats and half to Republicans. And if you did, I think you would probably end up with Democrats at 51 or 52. But the AI bans are pretty significant. I mean, it would not surprise me if Republicans had 52 Senate seats. It wouldn't surprise me if Democrats ended up with 53 or 54, but I think the median number is probably 51 or 52 is what I would think.
Sarah Spreitzer: I saw a prediction today about a 50/50 divided Senate with Vice President JD Vance being able to cast the deciding vote in the Senate. And I'm less bullish on the Senate. I think a lot of the races that are being considered toss-ups, I will be very surprised. And what I think is going to be really interesting is looking at the races where the president has weighed in, whether that helps or hurts the candidate. So the Texas race, I'm going to be watching really closely just to see, if Paxton wins, how much does he win by? Is it a really close race? Because that was a handpicked candidate by the president, he's been down to Texas a lot campaigning for the candidate, and it seems to be a pretty close race. But given that it's Texas, I don't think they'll flip the seat.
Mushtaq Gunja: Well, look, our baseline understanding should be that it's a significant uphill battle. I remember Jon and I having a conversation back in March in which he basically gave the Dems a 2% chance of taking back the Senate. The map is really hard because the map goes through significant red states. So the president won Alaska by 13.1 points. He won Texas by 13.7 points. He won Kansas by 16.1 points, Iowa by 13.3 points. So I mean, when you're looking at those margins in '24, I mean, you've got to have a 13-point shift, right?
Sarah Spreitzer: Yeah. And it's only a third of the Senate that's up for reelection. So-
Mushtaq Gunja: That's right.
Sarah Spreitzer: ... you're not going to be able to get to that 60-vote threshold that you really need to pass things with a filibuster-free cloture rule.
Mushtaq Gunja: Well, Sarah, I want to ask you about that in just one second, but I will note, I mean, getting to 50, or getting to 51 if you're a Democrat or getting to 50 if you're a Republican, does have pretty significant implications. I mean, the subpoena power comes with the majority, the ability to confirm judges comes with the majority. So I mean, it's a big deal which party will control the Senate. Now from a filibuster perspective and from passing actual legislation, Sarah, I'm curious, what are you looking for? It's very early. We don't even know. But from an education perspective, what are you watching for in these midterms?
Sarah Spreitzer: I think you're right, Mushtaq, it's not going to be a huge shift in the policy environment. The White House still has a very clear policy agenda that they're going to continue to push forward. Oversight is going to be the name of the game, especially if the House flips. There's going to be a lot of hearings, a lot of questions to Trump officials and administration folks about what's going, like on for the Department of Education, I think questions about the inter-agency agreements, questions about the secretary's ability to dismantle the US Department of Education. In other committees, likely questions about corruption and access to the president.
Current minority leader Jeffries has been pretty silent about impeachment possibilities for the president, but I think that there will be efforts to call up Secretary of Defense Pete Hegseth to answer for things. And I think that that means that there's just not much legislation being passed, and you still have a White House that has the ability to sign legislation or to send it back. And so it's not like there's going to be this complete shift in the policy agenda. What I'm watching more closely is how the election plays out for the White House, and if Republicans feel a bit less tied to the White House agenda, if voters are not happy with the DOGE efforts or the efforts to dismantle the federal government, will there be more pushback by congressional Republicans to reassert congressional authority on spending, on grant terminations, on things like that? That's really what I'm watching, but I think it's going to be a lot of hearings and not much legislation being passed.
Mushtaq Gunja: I think that's probably right. And we haven't uttered the term, I don't think yet this year, lame duck, but [inaudible 00:14:45] it's coming as, I think, year seven and eight of any two-term presidency over the last 50 years have been just littered with investigations, the things that have happened over the last six. I think that almost regardless whether the Democrats take power or not, I feel like we're going to be seeing some of that. And Republicans starting to try to think about what a forward-looking agenda looks like past Donald Trump, and Democrats of course as well as we enter '28 and '29. And that leads us maybe to artificial intelligence.
Sarah Spreitzer: Or as you referred to it, super intelligence, which has been the name given to it by the Trump administration. And I will note, all of the federal agencies are now referring to it as super intelligence. So I'll be interested in listening to your conversation with President LeBlanc to find out, do you refer to it as artificial intelligence or super intelligence?
Mushtaq Gunja: Well, I will say I'm happy to let the administration use their preferred pronouns, but I'm going to go ahead and use artificial intelligence and that's-
Sarah Spreitzer: Nice one.
Mushtaq Gunja: ... what we've been doing. I'll just note two things. One, this conversation with President LeBlanc is the first of two conversations around artificial intelligence. Coming up in short order, we will also have a podcast that we'll publish in the coming weeks with Studiosity CEO Mike Larsen about what he's seeing in the artificial intelligence space. And so that's happening. Also, if you were at all interested in this election conversation, we are going to be recording a live podcast the day after the election on Wednesday, November 4th. There's a link on the show page to register. Please come join us. I guess I've put some numbers out so Sarah can-
Sarah Spreitzer: Oh my God, Mushtaq is so excited about this episode, folks. And he's just going to come with all of the numbers. There's going to be so much paper. And I'm sure Jon will have his own thoughts and you guys can just nerd out about the election results.
Mushtaq Gunja: If Jon decides to come back.
Sarah Spreitzer: Yeah, if he decides to come back.
Mushtaq Gunja: Only kidding, everybody. We know that he will be back. There's no way he's missing this one. He loves a good election too. Okay, friends, we will be right back with our conversation with President LeBlanc. And we're back with our special guest, Paul LeBlanc. Paul is probably no stranger to this audience, former president of Southern New Hampshire University, the current CEO of Untamed Intelligence, a visiting scholar, special advisor at the Harvard Graduate School of Higher Education... That's not true. The Harvard Graduate School of Education, yes, Paul?
Paul LeBlanc: It is. You got it.
Mushtaq Gunja: And you would think I would know that because I used to play cards with a bunch of folks over at the School of Education, but most importantly for these purposes, author of the new book, Reclaiming Purpose: The University in an AI World, Paul, welcome to the podcast.
Paul LeBlanc: It's so great to be with you, Mushtaq.
Mushtaq Gunja: And Jon and I had the pleasure of reading your book in advance. Thank you for the advanced copy, Paul. And we can't wait to dive in. Jon, anything you want to say?
Jon Fansmith: Yeah. No, I mean, I am, as a government relations person, always a little hesitant to talk about technology issues. The Ted Stevens comments on the internet are burned into my brain, and I am acutely aware, Paul, that this is not my area of expertise. So I'm going to ask a little grace from you if I make a misstep. But first and foremost, table setting, your book is about repurposing higher education in the age of AI. Since you wrote your book, AI's gotten a little bit of attention recently, including some prophecies that we may end civilization within 10 years. I think we have tendency, I've seen you've identified this before, there's two basically very strong camps that approach AI as a technology, some especially maybe in some more traditional parts of higher ed that see this as disruptive and a tool for cheating and very damaging and maybe going to lead to the end of civilization as we know it. And then another school obviously that is fully embracing the technology and thinking about what that means and sees endless possibilities coming from it. Where would you put yourself in this spectrum? I think I could guess which side maybe you shade to, but I'd love to hear it from you.
Paul LeBlanc: Yeah, thank you, Jon, for the question. Thanks for having me. So I'm very much in the camp that says the obvious, which is AI is going to fundamentally change our world, probably the most profound change in our lifetimes. I think it is the kind of technological revolution that reinvents the world. We've had those before, but this will be, I think, incredibly powerful. And I have deep fears about how hard that transition will be. Optimist in the end, because I think it will give us the power and the tools to solve really intractable problems, like we're losing the race against climate change. I think maybe our only hope is the kinds of discoveries that will become possible, the kinds of things we're seeing in medicine today. But I think that we're just not prepared for what this transition looks like. We're not having an important conversation about how to cushion people who will be adversely affected in this, nevermind protecting us from annihilation, but we can go back to that question.
And in some ways, what we also have to be doing is carving out the space today to ask the question, what does a good life look like in this new world order? What do we want it to be? It's the question that FDR was asking way back in 1935 in the darkest period of time you could imagine. If you had said to people in 1935, "In just 20 years, America will be thoroughly reinvented," they would've looked around and said, "You must be crazy. What around you convinces you of that?" But in fact, we had to go through terribly, we had to get through the earthquake. The earthquake was World War II, it was the Depression, but FDR didn't wait until it was behind him to ask those questions. He started asking those questions early. And I think we have to start asking those questions. We have a rough road ahead of us, but my God, what amazing opportunity we have if we can get this right. And I know why we spend a lot of time worried about how it could go wrong. Let's carve out a little bit of time to ask the question, what does it look like if we get it right?
Jon Fansmith: So I'm going to just ask one follow-up too, because the reference to FDR, the New Deal, massive government involvement in every aspect of society. I think one of the other debates we're having right now nationally is what is the appropriate role of government around AI? And you see the companies themselves going forward saying, "We have concerns about the pace of development, what that looks like. We would like more regulation." Certainly our administration has said the opposite, that we don't want to stifle the growth of the technology, the advancement of AI, and certainly framing in terms of a competitive race against China and other countries. Well, I've got you on the big-picture AI-thinker space, give me your sense of what is actually advantageous. Is there an aggressive role for the federal government in harnessing and directing this technology to build that future life that you were envisioning, or is this something where a lot of other technologies have grown without that sort of support and direction, and where would you fall on that spectrum?
Paul LeBlanc: Yeah, so absolutely there is a critical role for the state in terms of getting this right. It inevitably only kicks in when we've had enough of a disaster that the citizenry will no longer stand for that idleness, which we're seeing right now. But bad actors possessing the power of these tools, that's a scary proposition. It could take other forms, but we know historically it always does, and historically the state always steps in. But here's the thing, it's never the state as it was constituted for the old world order. It's the state re-imagined. So if you think of the new world order that followed World War II when FDR convened the committee for economic development during the war, they start to ask the question, "How do we never have this happen again?" And they did things like the World Bank, the WTO, the Marshall Fund, all of those things were conceived in the middle of the chaos.
Mushtaq Gunja: I think you're on this question of the good life, because in reading your book, I really liked how you really centered what we were going toward, what we should be going toward. Can you spend a little bit more time talking about what you see as the good life? And then I want to turn us to what this means for our institutions and our universities.
Paul LeBlanc: Sure. So I think an industry that I think we all love, higher education, much of our sector has allowed itself to be rendered a sophisticated job-training program. We get measured by how well we do in placing people in jobs, we have gainful employment to say... And it makes sense, and that's not going to go away, but that should be our floor, not our ceiling. So as educators, we also want much more from education, and I'm talking about K-12 as well as post-secondary. And what we want, what we often say, and if you ask parents, what they will say is they want children who flourish in life. They want them to be healthy. They want them to have a moral compass, to have a sense of values, to have the critical tools to evaluate the hard decisions that we all face in life, and to make better decisions perhaps than they would with otherwise. As a state, we want a good citizenry, we want a sense of civic duty, all of those things. And I think we have lost so much of that. Much of higher education shies away from talking about values and morality. There are exceptions, and I talk about them in the book.
So if I want to be an engineer, I can be an engineer in many institutions, but if I'm an engineer at a really fine Jesuit university, like the one in which you teach, Mushtaq, I'm going to be an engineer within a framework where I'm also going to be asked to confront big questions about what it means to be a human being in the world, what do we owe each other? What do we owe the world and so on and so on? Formation is such a critical word in Catholic faith tradition, how do we form our moral and ethical compass? I could also then choose to be an engineer at West Point, and I will have a different set of values around service and duty and honor and leadership and so on. I could be an engineer at Morehouse, and I'll know absolutely I'll come out as an engineer, but I'll also come out a Morehouse man, and I will have a very clear sense of what it means to walk through the world as a Morehouse man.
I think these are very powerful examples of how we have to reclaim that ground that says, "Look, we are about preparing people for our world." And by the way, we should talk about how do you prepare people for a workforce that we don't understand the contours of, but I should also help people think about their embodied selves, the fact that they live in a physical body in the world, and that means something.
It actually ties to what differentiates our intelligence as well. I should educate for people for their ensouled selves, if I could use that phrase, and people sometimes roll their eyes a little bit. And you can frame that in the way that's most comfortable. So the Greeks would say psyche, Christians would say soul, Jung would say awe, that ability to feel awe, to be lifted outside of oneself. These are epiphany-like moments. And also values and morality. And I get why we might shy away or be afraid of talking about morality, but I'm not talking about giving you a dogmatic answer to that question. I'm talking about giving you the tools and the better questions that allow you to answer that for yourself.
That's a version of education, if you describe it to parents, like, "Yeah, exactly." If you describe it to employers actually, they say, "Yeah, absolutely." If you call it what it is, which is liberal arts, everyone runs screaming in the other direction. So I think we have to think about how do we reclaim that? And just as government can't step back in looking the way it did, I would think the liberal arts have possible for a very bright future, possibility for a very bright future, I don't think it can look like the liberal arts as it's been constituted. I think it has to be re-imagined as well.
Jon Fansmith: So let me talk a little bit more about what this transition looks like. We've mentioned, and you certainly mentioned, there's no one type of institution, incredibly diverse higher education environment here in the US and certainly internationally, roughly 6,000 Title IV participating schools, 21 million post-secondary students. Help me wrap my mind around, you've talked about the care economy, the idea of refocusing the institution to be very individual focused. Something, and we talked about this a little bit before, that resonated with me as a former Montessori kid, this idea of learning at your pace across a range of subjects, it just makes a lot of sense to me. It also seems very labor-intensive. And we constantly are dealing with issues with institutions that are curtailing what they offer, trying to balance student-support services against maintaining a course or a program or doing something else.
How feasible is a shift to that very individualized instruction, individualized attention, really a strong student-support network that's built around this, given all the other constraints we see in the higher education space? And you talked a little bit about the disruption. Is that where we get to that after a whole lot of disruption, or is this something that an institution sitting here now listening to this podcast could say, "I can chart a course there"? Walk me through a little bit, best case scenario in that regard.
Paul LeBlanc: Yeah, so I don't think we can overstate what a paradigm shift it is to finally achieve the goal of personalized N of one, personalized learning. The idea that I can utterly fine tune your learning experience to optimize it for you, Jon, and if you and Mushtaq and I were in the same class, we might all master, we would all master the same learning outcomes, but we might have three very different pathways to get there. So think about what a traditional education looks like today. It is still by and large a one-size-fits-all industrial-age model built for scale and repeatability and so on and so on. We don't have to live at that. Most of the world doesn't live in that model any longer. It's a legacy paradigm.
Now we have a technology that allows me to be available to you 24/7. If you are kid who is studying at 11 o'clock at night because those are your best hours, that system is available to you, adjust to your reading level. If we integrate with your wellness device, we can adjust the learning based on how much sleep you've had the last... It's pretty amazing what we can build. Tons of complicated questions there, by the way, we can come back to. But I do think that's the fundamental paradigm shift. So what that allows us to do, if knowledge transfer, that part of what a higher ed or what a faculty member does can be done much better on these platforms, it invites us to think about then what do we do with classroom time? What do we do with human time, with faculty and peer-to-peer time?
So I like the example of Kelly Anne Miller who teaches physics here at Harvard in the School of Engineering. She did a side-by-side test where she taught her traditional Intro to Physics class in the way she always has, in class, discussions, taking questions up at the board, blah, blah, blah. And then she did a number of sections where she ran students through a very different model where all of the teaching happened on a platform called Collage outside of class, an AI-powered platform. And her class time was all group work in Teams on real-world projects where she was circulating around the room, relational, students holding each other accountable for the learning. She had a 62% learning gain from the old model to the new model.
And I talk about it in the book, I call it the genius TA or the flipped classroom on steroids, where our value as faculty and as educators is not a knowledge transfer. I think it's in the way we curate community, we curate the classroom as intentionally relational space. And relationship, we know this is at the heart of all good learning, a really great learning. I also say, I stole this from a colleague here, but if you look at a toothpaste container, lots of ingredients. Education has lots of ingredients, but toothpaste always has one thing called an active ingredient. And without it, the other parts don't matter, they don't work. The active ingredient in really good learning models is always relationship. And it's also grounded in good psychology, and we know this. I wrote a book called Broken, and the first chapter is called Mattering. You can't transform people if they don't feel like they matter to you.
So I think when we go to your question, Jon, about how does an institution move there, this is such a big shift, because what it's asking faculty to do is to say, "Recognize that your value is not so much a knowledge transfer. You'll orchestrate that, you'll choose the platform, you'll do that stuff, but you now have to be engaged in relationship, and you have an invitation." And the best faculty have always done this. The reality is that if I were to ask either of you to name the faculty who were most impactful in your life, you would do that without hesitation, their names would be at the tip of your tongue, and you might remember them because they were good teachers, but I bet you remember them because you felt like you mattered to them, that they had a relationship with you, that they held you to a higher standard, that they helped you dream bigger dreams for yourself and so on.
So I think that's the way that institutions will look. So some institutions are there. I would argue that small institutions, the classic liberal arts colleges that don't get distracted by a whole bunch of other things tend to be that, students don't get lost, their faculty know them. I think in large institutions, you can have departments and even schools that achieve some modicum of that. Though if you talk to students in very large institutions and ask how many of them have a very close relationship with a faculty member of the kind I described, the answer is depressing. It's not very many. They get a lot of other great benefits, but it's often not that.
And how you get there as an institution is the hard question. It's easier to build a house than a renovate a house. So that renovation work, it could happen through crisis. Arthur would be the first person to say Brandeis could not survive in the old model, he's reinventing it. And I don't like the phrase never waste a crisis, but it is in moments of crisis that people get shaken out of their [inaudible 00:34:03]. Can't be saying, "What about your future," you start doing things differently. So I think there's that. I think we'll see new providers. It's very hard to identify very many new providers who then became dominant players, but they often perform a key function, which is they model or prototype things that the incumbent system can use and start to adopt.
Mushtaq Gunja: I'm thinking a little bit about what else colleges are going to need to do to reinvent themselves in an AI world and switching what faculty are doing in the classroom and how they're interacting with students, I think a huge one, absolutely. But I think one of the things that also probably needs to happen is some other and new way of being able to assess how well some of these skills are being translated, maybe not skills, competency, something. Because I think for a long time, especially our smaller liberal arts schools, institutions that were a little bit less focused on maybe getting students that first job with skills, but we're trying to teach students how to think for the long term, all of those sorts of ideas. I think we've done a pretty poor job as an industry of actually measuring those skills. It seems to me like this might be a time in which we might need to double down and do even more figuring out the right ways to be able to assess. Any reactions to that, Paul?
Paul LeBlanc: Yeah, I mean, I think I've argued for a while now back when I wrote Students First, that assessment's been broken in higher education, and I think AI has put the nail in that coffin. So right now we have a crisis of assessment in which we know students are using AI all the time in their assessments and their work. We've seen this enormous erosion of trust on the part of faculty who feel like, "My students are all cheating," because that's what it feels like. And I would argue that it's been a long time since we... I should rephrase this a little bit. Much of higher education doesn't actually measure genuine student learning. We measure other things by and large, the ability to cram for a test, and you forget two thirds of what you learned for that test a week later and so on.
So I've been a long proponent for a competency-based education, which inherently gets you into performance-based or authentic assessment. That is, it's not just what you know, but what can you do with what you know and can you demonstrate your mastery of that skill or capacity? So I do think that AI is going to force us to move and to confront this crisis of assessment. And also because AI changes the game in the sense that it's not so much about what you know anymore, because in some ways you can know anything. You have this super partner next to you that you can ask the questions, "Walk me through, help me understand," so on and so on. But what can you do with what you know?
And the knowledge of the tool itself is one of those questions. What can you do with AI? Can you simply get things done or can you ask better questions? Can you drive breakthrough thinking? Can you help us do work in new and better ways? Colleague here, Chris Dede, who retired from the grad school of ed, he has a great phrase, which is, "Don't use AI to do things better. Use AI to do better things. Let's ask better questions." So there's some really exciting, I think, prospects here. So one is I look at something like Dewey Learn, that's spelled D-E-W-E-Y, like John Dewey. So this is created by Luyen Chou and his colleagues. He's the former chief product officer at 2U. And it takes video and audio and allows you to look at performance. So think performance-based assessment again, but gives the student feedback in real time against whatever rubric the instructor has built into the AI, "So I want this performance to be based on X."
So let me give a couple of examples. It's being used at Cordon Bleu. Lots of practice kitchens have cameras right over the stove top to help students practice their culinary skills, their knife skills, "These onions look wrong, what's going on here?" And they're getting that active coaching in real time when it isn't being graded, when it's not high stakes, so they get better for those moments. I've been talking to the folks at Berkeley College of Music where I do a little bit of advising, and saying, "Look at a student who's practicing an instrument may get X amount of time at the end of the week with this master faculty member, but what if during the week they could train a camera on their guitar work, for example, and get that particular instructor's feedback because the AI has been trained on that?" So when they do five days later get to sit down with that human being, they're at a much better level and they're having a different conversation than they would've had if they had not had that feedback.
So think about nursing students, the simulation labs for student nurses all are equipped with cameras and audio and the ability to give nurses feedback on their performance as a way of maybe eventually of reducing the amount of times we require for clinicals, which is really a poor measure, just literally the amount of hours, versus, "Can you actually do the thing we're asking you to do?" And this can extend to soft skills or enduring skills, whatever preferred term you'd like to use.
So imagine, Mushtaq, that you were saying, "Oh, I want to work on my active listening skills or my negotiating skills," and I can put you in a simulation with other students and your assigned roles, but it's recorded, and then we can give you feedback at the end of that session, Hey, Mushtaq, there were three times in that meeting when your active listening skills are amazing. Do you want to review them?" "Yeah, of course I do." Like, "Oh yeah, great. But there are also two times you're off your game. Let's take a look at those." You watch it, "Mushtaq, do you know what happened here?" Maybe you say, "Oh, I was so excited with my next idea that I didn't listen to Jon. I just barreled right over his comment. I didn't listen, I didn't respond." Or you might say, "No, I though I did pretty well." Well, now we have a teaching moment. And that teaching moment records and rewrite.
So these are incredibly powerful assessment tools that we can use. We can scale oral assessment now. So this was always one of the critical problems, is it's expensive, it's not very scalable. We can scale it now with AI, and we can have audits of it and transcripts so instructors can take a look at it. So there's just, I think, incredible possibilities. But what we see a lot of right now is trying to preserve old assessment practices with things like AI bans and so on and so on.
And it's important to remember we still need to teach hard things. So we want to be really careful about the uses of AI that drives students towards more practice of hard things and more practice of relationship. And when you see AI that actually displaces relationship or eases the friction that we are no longer in the zone of productive struggle, not a great use of AI. So let me go back to the Kelly Anne Miller physics class. If you think about it, her use of that platform, that AI, allows her to then use her class for more relationship, working in teams, peers, she's moving through the room, and more practice. They're actually working on real-world problems. They're doing the work that will be expected of them. It's performance-based assessment. It's a beautiful illustration of what it looks like to get it right.
Jon Fansmith: So let me pick up on that a little bit too, because you talked about Kelly Anne Miller. I got the name right, correct?
Paul LeBlanc: Yeah, Kelly Anne Miller.
Jon Fansmith: Right. And within the classroom, students working together in small teams, common element, of course, it is students together in a physical space, sharing that space. We also have this model of constant assessment and feedback that is getting more capable, more individualized, more functional in terms of giving real-time feedback that meaningfully improves your skills or your learning process, and then an individualized educational method. Do you see those two things as in some ways inherently at odds, that we have students who come to campuses now, you talked about it, and say, "I want to go get a job. I want to do the shortest period of time for the least amount of money that will put me in the position to do the work I want to do at the best level of compensation." A lot of these other elements, they would probably strip out, and we've seen at times people who have sought educational models that try to reduce education to the barest bones to deliver that.
I guess I'm trying to grapple a little bit with attention in terms of the idea of the skills that are how you relate to and work with other people versus an educational model that says, "Well, we can deliver this skillset to you as an individual quickly, efficiently, maybe without a physical campus, maybe without classrooms and cafeterias and athletics and all the other things that are the trappings of higher education as we understand it right now." Do you see these core elements of historic, I hate saying that, higher education carrying through because they serve a purpose, or do you think we're going to start shedding a large part of what that experience is?
Paul LeBlanc: I think that the campus experience you describe is not of value to a lot of learners today. So when I was leading Southern New Hampshire University, my 32-year-old with three kids, a full-time job, two tours in Iraq were like, "What am I doing with the cafeteria?" Those are all part of what I would call the coming-of-age experience that classic residential campuses offer. My 32-year-old that I just described had all the coming of age he could handle. He needed a job, he needed a bunch of skills. And what we know as educators is he needs more than that. So skills may get you the job, but the reality is that human skills or durable skills that get you promotions, keep you in the job, allow advancement and so on. We've got to be able to do both, and we have to help students see both. I'm distrustful of providers and models that try to do one without the other. We've got to think about how we do both at the same time.
So I'm not sure I'm answering your question exactly, Jon. I think there'll always be space for that coming-of-age job. I don't know if I think it's a four-year task. The reality is that many students, when they arrive, like, "Hey, hallelujah, I'm getting up from under my parents' roof. I'm here for four years. Sign me up. This is like winning the lottery." Talk to them in their junior year, and they've got enough of what they wanted in that sense of coming of age, and they start to get itchy. They move off campus with their friends, they start valuing their internships more. By their senior year, they're itching to get out. So we already know that the four-year model is a suspect one, but I do think that desire for that phase of growing up, for those who have the good fortune to afford it, will remain something.
Remember also that universities do many things. So if you get into a place like Harvard where I'm sitting today, I'm looking to teach here at the graduates class, but part of what you get if you go to Harvard or Georgetown where Mushtaq is, is a value-added network. Got nothing to do with your skills. It's, "I'm in the club." I'm going in for a job interview and the person on the other side of the desk says, "Oh, I was at Harvard too. What house did you live in?" And all of a sudden you're connected.
These are important things. One of the things I was worried about at SNHU is we couldn't offer a value-added network of that kind. That's a value-added network of access and privilege and power, which our students didn't have. We started to try to build a value-added network, which was at least connection, like, "Hey, if I have a student who's graduating from this program and is applying to Fidelity, we tend to have a lot of SNHU graduates at Fidelity, can I connect them with someone inside who could give them good counsel on how to come into that interview more prepared?" That was really valuable.
So I think we have to think about the many jobs that we ask higher education to do, and then think about what's valuable. But here's what I would say, I want to go back to, and it's going to sound a little fuzzy or kumbaya, but I think as we think about the reinvention question that you asked, particularly in the age of AI, particularly in an age where we already feel much more dehumanized and fragmented than we once did, I think as educators and as institutions, we have to ask, what are the most impactful, important human relationships in this equation that we offer, in this thing that we offer to prospective students? And how do we identify it? How do we measure it so we have reassurance about it? And then how do we hold that ground sacred and protect it from all of the other forces that shape the business model?
So I think there's a lot of what we ask students to do and pay for that actually doesn't have much impact. And I do think there are models of the future where we might actually see more, not fewer faculty, but honestly fewer administrative staff, because a lot of administrative work can now be done really well by AI. That's a painful transition. That's not going to be easy. And it's the kind of workforce displacement that every industry is now worrying about and thinking about. But I mean, I think what we want to make sure we do is protect the most important relational and human moments in our learning equation, and we know they're critical.
Mushtaq Gunja: Well, hopefully some of those college administrators have some of those durable skills that will allow them to be able to move to the next sets of jobs, because some of this is coming regardless of whether we want it to come or not. Paul, you made an analogy or you referenced using AI to ask not just to have your job be done better, but to ask better questions in the first place. [inaudible 00:48:10].
Paul LeBlanc: Yeah.
Mushtaq Gunja: And we have a lot of senior administrators listening to this podcast, Paul. What's the question, maybe boil it down to one or two, that you would have the folks on this podcast that are listening ask themselves as they enter the next year or two where disruption is certainly coming?
Paul LeBlanc: Yeah. So one, I think we're going to see a very spiky transition. So this is not going to happen without a lot of pushback and backlash, which we already see in America. It was interesting, Wall Street Journal last week, I think, reported two-thirds of Americans would like to see us hit the pause button on AI and stop building data centers. Two-thirds, it's bipartisan. And if you think that nursing licensing boards are going to allow nursing to be reinvented overnight by OpenAI, ain't happening. I think it's a good thing, and we see just at the state level, lots of policy, lots of bills being proposed to control this, and, "Where's the human in the loop?" So we have this messy transition over the next five and more years. And it's probably a good thing because we need to buy ourselves some time to be more thoughtful about this.
Now to go to your question, I would say, here are the following questions when I sit down with presidents or chancellors or boards, what's your theory of the future of work? You don't have to have mine, but you have to have some theory of the future of work, because you can't answer the question, "How do I prepare my graduates?" if you don't know what you're moving them out into, you don't have a theory about that. And what we see emerging is that, and I wrote a piece for the Washington Monthly this summer for their rankings issue, which was don't bet on a job, bet on a sector. So if I'm the dean of the School of Engineering, or I pick any school, school of education or healthcare, health sciences, I have to think less about what does it mean to be specifically a PA or a nurse or a physical therapist, but what are the higher level skills that I need to prepare my graduates? And then the tools of resiliency, because they're going to navigate tons of change. Even the most conservative view of the future says we're all going to navigate tons of change in the nature of our jobs. So I have to have the higher-level skills.
So if your son or daughter says, "I really, really want to be a graphic designer," talk about what it means to be a creative, because graphic design, who knows? But they're probably always going to be creative jobs, I believe. What does it mean to operate in that sector? So that's one. So what is your future of work? I actually think that we're going to see a migration of talent into human-first jobs, or what I call care-economy jobs. Allison Pugh at Hopkins calls them relational jobs, where the human doing the job is more important than the technology and the AI that supports the job. We could talk about that if you like.
Second question I would ask, again, people leading organizations is what's the vision of this that you're bringing to your stakeholders? Do you have a vision for how the university will look in the future with AI? And if you don't, and it's very hard to have one right now, can you at least articulate a vision for how you'll get there? Do you have a vision for the journey? And I think the journey has to have a set of radical commitments. And I love this phrase from some of the change management literature. So Brown in some ways made radical commitments and said, "We're going to double down on what makes Brown Brown, and this is what we believe." So it's a radical commitment to our mission and our values, that would be one. If I were leading today, it's easy for me to say that because I'm not, but if I were leading, I'd say, "I can guarantee you your job is likely to look different, but I will also guarantee you, I'll make this promise to you, that I'm going to do everything possible to make sure you have a job that you find impactful and meaningful. And we're going to figure this out together. I want to be hip to hip with you." There's no top-down voice from on high saying, "This is how we're doing AI." I think it's a mistake.
And then I think the third is when you look at deploying AI in your institution, institutional leaders are facing enormous pressure from their boards to do so. I think there are a series of thoughtful questions that have to be asked about that, and they include things like, "Are you transparent? Do people know how you're using AI?" There's seen AI and unseen AI. So seen AI is when you go to ChatGPT to ask a question. Unseen AI is when you turn on Netflix and all your recommendations are there for you. It's happening with AI, you just don't get to see it. And then is your data architecture in order, including increasingly security, cybersecurity protections, data protection? Because we are moving into a crazy-scary time in terms of that with the new models. I would ask the question of, do you have clear policy that students understand across the board? Students are so frustrated that they get rewarded in one class and then punished for the exact same use of AI in another class.
So think about policy as three tiers, fixed, flexible, and free. What's true for everybody? It's fixed. What's flexible and who do you give that to? So the policies of the financial aid department may look very different than the policies of the athletic department. And then what's free? You're a faculty member, these are the decisions you get to make, but they don't get to countermand the ones below it. It's a scaffolded approach. So there's just lots of these really important deployment questions. But Allison Bailey, who's the head of talent and culture at BCG, she points out that their work with organizations, lots of organizations, it comes down to 10% data, 15% technology, 75% people and culture. Your work is not really on the technology side at the end. Your work is on your people, is on the culture that you've built and that you're trying to lead and shape.
The other thing I would say is that, as I said earlier, it's harder to renovate a house than build a house, so build a shed. In other words, start renovating the house, do the things you need to do that we just talked about, but have a small team that's sitting off on the side that's given the permission to build dramatically new models. So if you have an incredibly great MBA program, have a small team and challenge them and say, "Hey, invent the AI-powered MBA program to put our MBA program out of business. I'm not going to give you a lot of money and not give you a lot of people because you shouldn't need either in a world of AI, but be really imaginative and bold. No holds barred, what would you do?" So you get the disruptive innovation work on the margins even as you're staying true to the core.
Sorry, that was a long answer, what people leading institutions should do, but I talk to a lot of them, and some of them are doing great work. Look at UC San Diego, he's doing incredible work there, and they create these little play sandboxes that any faculty or staff member can come in and use AI to solve a problem that they care about. They call these citizen apps, and they do like 200 of them a year, and most of them never see the light of day, only that person uses them, but sometimes they invent something. But what are they doing? They're raising the water level of everybody. Everyone's getting familiarity, they're learning how to think about the tools. So there's some good examples of people moving out, I think, in interesting ways.
Mushtaq Gunja: Well, it might change-
Jon Fansmith: Yeah, we definitely have to have you back to talk about that too, and policy setting at institutions, which could be a very long conversation as well.
Paul LeBlanc: It's so hard. Yeah. And it's [inaudible 00:55:42] hard because the target is moving. The technology's racing along so fast that... And I think we have the complicated questions about anthropomorphism, there are complicated questions about workflow and learning design. I bet you, I was a writing teacher and I talked to a lot of writing teachers, and almost all of them would uniformly say, "I don't want my students using AI to write their drafts. They're not learning the thing." I would agree, they're not doing the hard work. But I say, "Do you encourage your students to use AI to generate topics, like, 'Give me 10 possibilities for an essay,' is that okay?" And most of them will say, "Yeah, that's okay. That's pre-work, and I don't mind if they're..."
And I talked to one of my colleagues, Tessa Forshaw here at Harvard, who's a cognitive scientist, and she said, "Oh, that's a terrible idea." I say, "Really? Why? It sounded okay to me." She goes, "Because when you think about creativity, creativity comes when two ideas are furthest away from each other inside your semantic web, inside your neurological web." And if you've ever had that experience of looking at some creative piece and you say, "Oh my God, who would've ever thought about putting those things together in that way?" You're exactly capturing what she describes. AI does the opposite. It goes to what's the most likely next thing, which is why AI writing, AI art, AI music, it's all meh. It's like, "It's okay, but it doesn't take your breath away. It doesn't feel original or creative. It doesn't feel right. It feels meh." The problem is we live in a culture where meh has become acceptable. A lot of our cultural artifacts are meh.
Anyway, I've gone on a tangent. My point is we really now have to start to spend a lot of time thinking about how do our brains work? How do we help students understand how their brains work? We use the term metacognition. No one else does, it's not a term that people resonate with. But if you said to someone, "Hey, is this stuffy toy with an AI agent good for my five-year-old?" It's like, "Let's talk about how your five-year-old's brain works and what isn't true yet." And what we would say, by the way, is, "No, it's not okay. It's a bad idea, and here's why." We need to do that with adults, we need to do that with workers. When we talk about designing workflows, let's say the back office within higher ed is like, "Oh, we're going to use AI to do these workflows, but we're going to have a human over the loop." That's the new phrase, not in the loop, but over the loop, monitoring. I think HBS has some new research that tells you what happens if you put humans over the loop, they take their hands off the wheel. It gets boring. It's like, "Hey, if that AI does this process 100 times perfectly, why the hell am I looking at it the 101st time?"
So it's designing a workflow that ostensibly is good because humans are in control. And you know what humans do? Our brains get lazy, like, "No, we're built for efficiency. I'm not looking at this anymore. I'm going to go shopping instead." So we have to think about designing workflows that keep people engaged and maintain our, the key, not all of them are important, but the key and the most important cognitive functions that we have.
Mushtaq Gunja: Well, I heard Jon say, "We're going to have to have you back." And I heard you, Paul, say, "Yes." So I think we're going to take you up on that. I know-
Paul LeBlanc: Well, thank you. I'd love to be back. It's always good to talk about these things.
Mushtaq Gunja: I'm going to speak for Jon and say we enjoyed this conversation immensely. You've given us a lot to think about. We'll have you back on in, well, maybe a year, year and a half, ask these same sets of questions again, I think, and see which answers we get that are the same, but I assume things are going to change quite a bit.
Jon Fansmith: If you enjoyed the show, please consider subscribing, rating, and leaving a review on your favorite podcast platform. Your feedback is important to us and it helps other policy wonks discover our show. Don't forget to follow ACE on social media to stay updated on upcoming episodes and other higher education content. You can find us on X, LinkedIn, and Instagram. And of course, if you have any questions, comments, or suggestions for future episodes, please feel free to reach out to us at podcast@acenet.edu. We love hearing from our listeners. And who knows, your input might inspire a future episode.