Live from Google: AI Is About People, Not Just Tech

Chapters

0:00 - Episode Trailer
2:20 - Intro
3:36 - Why Are Students Both Using, and Booing AI?
6:10 - How Should College Leaders Talk About AI?
10:05 - Is Inconsistency in AI Policy a Feature or a Bug?
12:52 - Using Data to Drive Faculty Discussions on AI
14:33 - Invest in People, Not Just Tools
16:33 - Helping Faculty Feel Safe to Try AI
20:24 - Can Colleges Afford AI In Times of Budget Crisis?
23:26 - Encouraging All Disciplines to Try AI
26:30 - Do Professors Trust Administrators on AI? 
30:14 - Sponsor Break
30:44 - AI’s Impact on the Value of the Degree
34:48 - How is AI Really Impacting the Job Market?
38:08 - The Future of Assessment
42:38 - Does Traditional Faculty Governance Work for AI?
45:46 - What Professors Do That AI Can’t
46:41 - When Industry Moves Faster Than Education
49:12 - The Importance of Faculty Development in AI
52:15 - Closing Thoughts on AI at Colleges

Relevant Links

“AI Is Routine for College Students, Despite Campus Limits,” Lumina | Gallup.

AI in Action video series at California State University System.

“Cal State students widely use AI tools, but mistrust results and fear job impact,” in EdSource.

Transcript

James Frazee

One of the quotes, open-ended comments that really stood out to me was a student who said very politely, ‘Please just tell us what you expect and be clear about it.’ Because there's such a high degree of variance between courses. They can go into their 9:00 class and be encouraged to use AI, and they could be expelled for doing the exact same thing in their ten o’clock class.

Jeff Selingo

Michael, that's the CIO at San Diego State describing one of their students who answered a survey about their experience with AI.

Michael Horn

And that's the elephant in the classroom on every campus that no one wants to reckon with.

We took Future U to Google in Sunnyvale to look at it.

Jeff Selingo

It's an all California panel.

The chancellor of UC San Diego, the chancellor of the California Community Colleges, 116 campuses, 2.2 million students in the system.

Michael Horn

And the CIO of San Diego State.

And I went and convinced Jeff that a patchwork of policies was a problem to be solved.

Sonya Christian

Is that a problem? To me, it's not a problem. If we had too much coherence in academia, that would be a problem.

Jeff Selingo

We get into what actually moves faculty.

James Frazee

We need to invest less in tools and more in people.

Michael Horn

Whether the degree survives all of this.

Pradeep Khosla

The best skill we need to have, we need to impart on our students, is the skill to acquire skills on their own.

Jeff Selingo

And whether California brings back the SAT.

Michael Horn

And then somebody said the thing that's been rattling around my head ever since.

James Frazee

The deliberation that is, you know, perceived as, you know, at a glacial pace is not a bug. It's a feature.

Michael Horn

It's Future U., live from Sunnyvale.

Let's go.

Sponsor

Bring the best of Google's secure research backed AI tools like Gemini for Education and Gemini Notebook to your institution. Elevate leaders, faculty, and students with purposeful technology, transform learning, and drive innovation. Learn more at edu.google.com.

Subscribe to Future U wherever you get your podcasts. And if you enjoy the show, share it with your friends so others can discover the conversations we're having about higher education.

Michael Horn

The goal for today's discussion is to first name some of the disagreements at play on campuses between students, faculty, administrators, frankly within and between the different disciplines, between different types of colleges and universities and the broader ecosystem and not pretend any of those as monolithic as we do so, about the role that AI should or should not play in the future of higher ed.

And then after that meaty discussion, we're gonna shift a little bit to, How do we bring the various constituencies along on the journey wherever that may lead?’ Which is vital given that the external need for AI skills is vast and fast moving.

So with that, Jeff, let's introduce our panel. Yeah.

Jeff Selingo

So we have an all-California panel today.

So please join me in welcoming first Sonya Christian, who's the chancellor of the California Community Colleges.

Pradeep Khosla, who is the chancellor of the University of California at San Diego.

And James Frazee, who is vice president for information technology and CIO at San Diego State. James.

So we're gonna get started with a quick lightning round first.

So recently, Lumina, the Lumina Foundation and Gallup came out with a survey that found 57% of college students use AI weekly in coursework.

Yet, as we all know, this spring, we had all these commencement speakers going viral at colleges and universities across the country for getting booed at commencements for praising AI.

So on one hand, you have students who use the tool daily, but yet they reject the celebration of it in the same breath.

What does that contradiction tell you with the students that you're dealing with?

Pradeep, let's start with you and we'll make this pretty quick.

Pradeep Khosla

Okay. So I think both are right.

To the student who's using the tool, it looks like an enhancement of their capabilities and amplification. Now they are able to do things that they would not have been able to do without the tool.

At the same time, they don't trust the tool in other people's hand. To them, the administrator who gets the tool is going to create deficiencies and job losses.

And I think what we need to appreciate is this generation that we are dealing with right now between ages like 15 to 30, they are more socially aware, socially conscious about issues that previous generations were not. They care about this.

So I think both are right.

Jeff Selingo

James.

James Frazee

Yeah. I think students are incredibly pragmatic. They see the value of AI, and they're also concerned about the impact on their learning. The impact as Pradeep mentioned on their job uncertainty and the narrowing of the entry point into the job market. And I think they're concerned about becoming over reliant on these tools.

So I think they see the value and they also have legitimate concerns and they hold those two things at the same time.

But I think they're very pragmatic.

Jeff Selingo

Sonya, how about … what are you seeing among your students?

Sonya Christian

Well, the same thing. You know, our students, they don't want AI to be like the hero of the story. They want the story to be about the human, the students, their future and AI being really a tool.

Truly your title for this podcast really captures the essence of what our students are looking for. You know, jobs, I mean climate issues, that social consciousness is front and center for them.

Jeff Selingo

So what should you do about this then in the roles that you play? So what should colleges and university leaders do?

Do they have to act any differently now going in as we kind of think about AI and its role in higher ed?

Are you thinking about how you present this and talk about it on campuses into different constituencies?

Pradeep Khosla

Actually I am. So I think we should not extol only the virtues of the technology and the tool. We should also bring up concerns and genuine concern for people and how this technology might shape humanity. Right?

So I'm chairing a UC wide steering committee right now in AI, and it's called AI for Humans and Humanity because I think the human being is gonna change and humanity is gonna change, and we need to be careful about how we deploy this technology.

So we have to talk about it on both sides.

James Frazee

Yeah. I think you know it gets down to trust and we need to make sure that our students feel like we trust them.

I worry about the erosion of the relationship between the faculty and the students when the students feel like they're being falsely accused of academic misconduct, when often they're confused.

You know, we did this big survey in 2023, and we've replicated it. In fact, just this last fall, we scaled that up to the entire California State University system. Over 80,000 students responded to the survey.

And one of the quotes, open ended comments that really stood out to me was a student who said very politely, ‘Please just tell us what you expect and be clear about it.’ Because there's such a high degree of variance between courses. They can go into their 9 o’clock class and be encouraged to use AI, and they could be expelled for doing the exact same thing in their 10 o’clock class. And that's causing a great deal of anxiety and stress among our students.

So, I think that element of trust is really important. They need to feel trusted.

And the faculty also need to feel trusted that they own the curriculum. This is their call. They have the freedom to choose.

And I think that puts a lot of onus on us as leaders to provide them with lots of examples of how they can use these tools effectively to get the students engaging in the content maybe more deeply than they might have otherwise.

Jeff Selingo

And so finally, Sonya, how are you thinking about this as a leader?

Sonya Christian

So where I sit is at the system level.

So we have 116 colleges, over 2.2 million students. We have students that come to us right from high school but we have a large population of returning adults and working adults. So it's a very diverse group of students.

We in fact have a student who routinely comes to our Board of Governors — it's a 17-member board appointed by the governor of California — comes every meeting and talks about data privacy and data security for the students.

So one of the very first things that we did was to adopt a set of principles and we kind of cutely spelt out the word H-U-M-A-N as humans. But the principles really talk about data privacy. It talks about having humans in the loop. It talks about algorithmic biases and how do you when you're doing procurement to make sure that we have those values that are built into our partnership with our edtech partners.

So that has been really as released and allowed the California Community College faculty members to create a space for them to experiment.

We also created a Digital Center for Innovation, Transformation and Equity to give permission for folks to be able to experiment within the principles that we have operationalized in our agreements with our edtech partners.

Michael Horn

So James, you went exactly where I want to go next, which is you mentioned the relationship between faculty and students around all of this.

The same Lumina Gallup data that Jeff was referencing is striking because the majority of students, as we mentioned, use AI weekly.

But just as many say faculty rules are unclear at the moment, they're certainly uneven.

And so I'm curious from each of you, is that incoherence that students are experiencing right now, is that a faculty problem? Is it a leadership problem or an honest reflection that frankly none of us knows where this is gonna end up and it's a moving target?

James, yours you sort of already gave a take on that.

So maybe, Sonya, let's start with you and then, Pradeep, after that. And then, James, if you wanna jump in on the back end, go for it.

Sonya Christian

I really think it's where it needs to be, the conversation.

So sort of the friction points, the debate, the discussions of what does academic integrity look like, right?

And students are asking for a certain set of norms while faculty are debating where does innovation and where do the guardrails fit together.

Is that a problem?

To me, it's not a problem. If we had too much coherence in academia, that would be a problem.

But in many ways, it's sort of a microcosm of the largest society. In fact, just several weeks ago, I got a long email from a faculty member who gave me a link to Dario Amodei's recent treatise that he put out about the risks of AI.

And what the point the faculty member was making was, you know, in 2024, he was really showcasing the virtue of AI and now there are all of the risks.

So I think the debate and the discussion is resulting in an evolution. And being engaged is really the trick to it, not trying to fix everything all at once.

Pradeep Khosla

So I think even before AI, right, people had closed-book exams. People had open book exams. People had take-home exams. People had no exams. I mean there was a whole bunch of variety of methods and tools that individual faculty were using to evaluate your learning in that course, right.

So I think AI is no different, except in some sense, we need to educate the students that there is an independent faculty member to evaluate your learning. And if they give you a set of rules, then those become the rules for that course. Those rules don't have to be the rules for all possible courses because academia is structured very differently. It's not a company where there's a set of policies uniformly applied, especially in the classroom.

James Frazee

Yeah. I think that my answer to your initial question is neither the faculty nor the leadership.

We've found that using data to really scaffold conversations among the faculty is the most important thing we can do. In fact, we have an initiative underway right now called Leading AI-Ready Programs, which is for our faculty department chairs and school directors and curriculum committee chairs. And in fact, we have a faculty member from San Diego State from our College of Business who's in the audience who's going through that experience right now.

Importantly, the faculty are being paid to go through this experience as an important sign of respect that their labor has value and that they're not in this alone.

And the idea is to provide them with a toolkit that they can use when having conversations with the members of their faculty in the fall so that they can talk about this patchwork of expectations that is causing a great deal of anxiety among their students within a particular discipline or domain, and bringing that back to the degree learning outcomes, the accreditation, expectations for their particular field.

And we're using data from that survey I mentioned earlier to really scaffold that so they get to see how their department compares to their college, how their department compares to the university writ large, how their department compares to our system, and that data is really essential to get them thinking about their students within their program, and they care deeply about their students and what they have to say.

So, this is a really important initiative that we have underway. We're not prescribing anything. We're just creating the conditions for them to have these conversations.

Jeff Selingo

James, let's stay with you for a second because I think we live in a pretty ambiguous world right now around AI. We're still trying to figure this out.

It is also a little bit of a cat and mouse game on campuses. Right? We're gonna do this, but not this, and we're always trying to catch up.

So in the early years so far, are there things that you have spent time and money on that you're now thinking, ‘Oh, that's not really paying off, and if we have to wave a magic wand, we should do this instead?’

Have you figured that out yet or is it still too ambiguous?

James Frazee

You know, I think the thing that is really clear is we need to invest less in tools and more in people. And it gets to the title of this conversation.

And we did another faculty development experience over the winter break where we recruited faculty from every one of our undergraduate degree programs. We've got over a hundred undergraduate degrees. And the idea was to again compensate them for thinking about how they could retrofit their curriculum with this new reality.

And one of the deliverables for that in order to get the funding was to create a reflective video talking about what they did, why they did it, and most importantly, what were the outcomes. And that's contributing to something we call the AI in Action video series because the faculty in biology want to hear from other biologists. They don't want to hear from historians. Right? So now we've got examples from across the continuum of the enterprise that could resonate with somebody about how they might use these tools within the fabric of their curriculum.

So I'd say invest more in that and make sure that those types of experiences are by faculty and for faculty. They're not being led by certainly not the administration. So, that has really served us well.

So, I'd say invest more in things like that and less in the tool set.

Michael Horn

So this starts to bring us to the conversation where we think about faculty roles changing, how they adopt, how they innovate, take risks, and so forth.

There's a lot of research, as you know, all of you, about how to help individuals and organizations feel comfortable to take risks. Some of the research suggests that pushback often that looks like obstruction is maybe closer to grief. It's sort of challenging their identity, their sense of how they accomplish the work.

My late mentor, Clay Christensen, used to like to say, there are no stupid people, that organizations, in fact, compel people to act in certain ways.

And the research is also blunt, and Pradeep, I wanna go to you on this, which is people won't take risks without a stable foundation, and yet you could argue right now faculty are being asked to experiment amid a lot of headwinds — research cuts, demographic changes, shifts in international enrollment.

How do you encourage faculty to feel safe to experiment with AI given that external environment?

Pradeep Khosla

So I'll talk about UC San Diego.

If you look at UC San Diego, about two thirds of our faculty are tenured. So at that point, telling me that you're not feeling safe is not a good idea. Okay? Of all the jobs in this country, that is the safest job in spite of all the cuts.

Now having said that, my view as a leader and administrator is I don't want to top down force anything on people. I want to show people what is possible. I want to let a thousand flowers bloom. And I want these smart people. I mean you don't become a great institution by hiring dummies. You become a great institution by hiring people who are self driven. Even if I say don't work, they'll come and work sixteen hours a day. So to show them what is possible, open up all the lanes, and make sure that I'm not censoring one way or the other. I'm not telling you what to do and I'm not telling you what not to do.

So we had, so like, you may not know this, but my PhD is from Carnegie Mellon '86 in robotics. So I was in this AI game at DARPA long before AI. I never thought ChatGPT would happen in my lifetime.

So when this was announced, that very day I sponsored 10 projects on campus, right, to look at admin and academic side.

Then about three months ago we held like two years later a conference of tell me what are you all doing in AI. Three hundred faculty showed up who I had no idea were working in AI. I had no idea how they were using people in humanities, people in history, people in social sciences.

And what I find is I think as leaders we need to trust the creativity of our faculty. They know what is right and what is wrong for themselves, and we need to be inclusive about the rights and wrongs. We cannot be that only one thing is right and the other thing is wrong because in this business…

So I'm very positive about where the faculty are going, how they're thinking about it, and I'm not a great believer in a top down data. Top down, I think we've to create a force field, but bottom up we need to let a thousand flowers bloom and let them be in a very wide force field or enabling field.

Michael Horn

Sonya, let me turn to you with this next question because you're overseeing a system, as you mentioned, 2.2 million students with arguably the most at stake, the students that you serve.

And I'm curious, does scarcity of resources make the equity case for AI stronger among faculty? Right? It's a way that they can amplify their efforts, staff efforts to better support students, or does any AI initiative at the moment feel like a luxury that's sort of financially irresponsible to invest against right now?

Is there real budget, in other words, to chase after these big initiatives that move against important goals for students?

Sonya Christian

And I love that you brought in the word equity because when you think of community colleges, that word is really embedded right in there.

So I actually was applying for my position as state chancellor in November 2022. And you know what happened in November 2022 is when LLM, ChatGPT, hit kind of the market in sort of a commodity fashion. And so I brought it into my cover letter, and I was given the position in February of 2023. And in July 2023, our board of governors for the entire system adopted AI in the future of teaching and learning as one of the three strategic directions. So they were all into this, which means if a governing board is all into it, it means resource allocation. Right?

So the way that … what one of the guiding principles was not to invest in heavy technology. So we didn't spend millions of dollars on technology. We invested it right from get go with our faculty.

And often when you think of innovation happening, at least my structure was it's a bell-shaped curve. You have those innovators at one tail that kind of pull the rest of the curve.

But with AI, we felt we needed to move quickly because the half life of AI and its evolution is shrinking as we speak. So we deployed a model that just we refunded faculty to get into our community colleges, and we created micro disruptions led by these talented individuals. So we invested in release time and we brought them together and this is where we came up with a set of principles to be anchor.

So we adopted what is called a tight loose model. So to allow a group of, you know, Pradeep talked about 300 faculty when we went out and said, you know, the chancellor's office is going to be doing AI fellows and we have 13 positions. We had over 200 people apply for it. And so we had to create a structure to support much more than 13.

So what ended up happening is the tight loose principle is we grounded ourselves in operationalizing what does data privacy look like? What is human in the loop look like in decision making? What does algorithmic discrimination look like operationally?

So once we had that conversation to kinda create a norming structure, these individuals were out there creating those micro-redesigns.

So the changes that you see happening in the California community colleges over the last two years has happened organically from departments, from academic senates, from other consultation council structures. And the only thing we needed to do is to guide and support and create the conditions for the creativity to thrive.

Jeff Selingo

So James, I'm curious about how universities create those conditions.

So Sonya just mentioned, right, coming up from the departments. You mentioned biologists want to hear from biologists, historians want to hear from historians.

But if you look at the Gallup Lumina survey, you see big differences in faculty opinions. College Board has also done this, faculty opinions by disciplines, right?

So how are you trying to encourage this evenness across campus? Like so that there aren't silos, so that it's not the just the business school and the STEM disciplines doing everything and everyone else is kind of being, you know, trying to come up behind those schools.

So how do you create these conditions to encourage different disciplines?

James Frazee

Yeah. I think it's doing things like we're hearing about at UC San Diego where you have events, forums that bring people together so that they can share in a very public setting the work they're doing. And that's how you can really not only recognize them, but celebrate them.

So, as part of that AI in Action video series, one of the things the faculty get is a letter from the provost thanking them for their contribution. And, this is a big deal, right? When they go for up for review for tenure and promotion, they put that in their package as one of their, you know, pieces of evidence that they're contributing to the university as part of the service expectation related to their work. And, I think those doing that kind of thing in a very formal way is really important.

The other thing, you know, it gets back to what Sonya was saying around values. You know, I think that people are looking for a silver bullet, and they'll say, oh, we need an AI policy. And, my answer to that is we have an AI policy. It's our academic integrity policy. And, if you change that policy with every new emerging technology, you know, the ink's not even gonna be dry on the policy before the technology has now evolved to a point where you haven't addressed every potential, you know, element that needs to be thought about.

But, when you have principles and when it's based on values, keeping students at the center. You know, promoting curiosity and creativity among the community, those things are really timeless, and you can keep going back to those principles.

The principles that we came up with were developed through a university senate group. There was a Delphi process. It took 18 months to come up with those draft generative AI principles.

But it was about you know, it's not about the product or the principles. It's about the process.

And just, you know, same thing can be said with the curriculum. You know, we're trying to encourage faculty to move away from these high stakes assessments and put more emphasis on the process, iteratively improving and getting feedback from others and really exercising human judgment as really a critical component of their pedagogy.

And again, it's about examples and giving them lots of examples.

Jeff Selingo

Okay. Let's wrap up this conversation about the faculty with kind of a lightning round around trust.

Do you think there's a lack of trust right now between administration and faculty on a lot of campuses. Do you think, and I know the faculty are not a monolith. Do you feel like the faculty kind of trust where you're going on this — how you're thinking about it?

Pradeep Khosla

So I don't know if I can say faculty overall trust where we are going, but I have to work on the principle that we as administrators are going to be as transparent as we can, and we're going to communicate, communicate, communicate. And it takes time to build trust. And in difficult financial times, it takes more time to build trust.

So we cannot give up just because I gave the lecture twice and nobody trusts me that they're not going to trust me. Faculty are super smart, hardworking, intelligent people, and they're all committed to the success of the organization. So I have to believe that I have to keep on communicating and overall the trust will keep on improving.

There's no notion of 100% trust versus distrust. There's a level of it. So it keeps on improving. It's a process. We've got to go through it.

Jeff Selingo

James.

James Frazee

Yeah. I Think you have to walk the talk. One of the things that we've done is we put together a generative AI privacy statement. And we did that not only for the students. So the students knew in … And it's short. It's a paragraph long. And it says in no uncertain terms, we're not accessing, monitoring, or tracking the inputs or outputs of your use of our enterprise tools.

And that's not only important for the students to hear, but it's also important for the faculty to hear. You know, they want to protect their intellectual property, and we encourage them to use these enterprise tools because we have commercial data protection, and they can rest assured that information that they're entering and the output that is being generated isn't being used to train those underlying large language models.

But that's a process of, you know, education with the community of, you know, if the product's free, you're the product. Right? You're paying for it with your ideas and your intellectual property. And also to do that in a very explicit way.

And in fact, that privacy statement we had one of our associated student government leaders do a short video talking about that privacy statement. And that's been woven into the fabric of an AI microcredential that's required of all of our first year students.

So, they get right when they enter the university exposed to how to use these tools responsibly and why not to use the free versions of the tools.

Jeff Selingo

Okay. Quickly, Sonya, trust among the faculty.

Sonya Christian

Yeah. So, I mean, it's a continuous conversation. And rather than framing it as sort of a bilateral relationship between administration and faculty, it's the entire ecosystem.

So, the California Community Colleges, our conversation, we have industry partners at the table. We've structured and designed over the last several years, fellowships where faculty members are working. I mean, we've got a great relationship with Google, for example. So when there are these different conversations that are happening, it's all of the collective conversations coming together.

We've also structured conversations where students are at the table. So students, faculty, there's also our classified staff who are at the table, you know, who are worried about their own jobs.

So I think creating the dynamic where we're really trying to represent the entire social structure when the discussion and not force it to be like a bilateral conversation.

It is an issue. It's the age of AI. We're all going to deal with it together.

Sponsor

Bring Google's secure research backed AI to your institution to accelerate research, streamline operations, and transform education. With Gemini for Education and Gemini Notebook integrated across disciplines, you can boost student engagement, improve academic outcomes, and personalize instruction. Empower your entire community to collaborate, solve complex problems, and prepare students for the future, all on a safe, scalable foundation. Discover more at edu.google.com.

Michael Horn

So we've talked about how faculty adopt AI.

There's another question lurking, which is, and some of you have alluded to it, which is, f AI can produce outputs or facsimiles of outputs that we thought only humans could, for many, it raises existential questions around the goal or the value of the degree, around how to prove what students are learning.

For some people, AI is actually an answer to a part of that question.

But, Sonya, I'm curious your perspective on this.

We do see that the value of the degree is under attack from lots of different directions at the moment. If a student can produce competent work with AI that used to require years of education, it seems that there's at least three possibilities.

One, that, yes, the degree's value's gone down.

Two, maybe the degree needs to move somewhere that AI can't reach.

Or three, perhaps we've misrepresented the real value, the intrinsic value of the degree itself.

How do you think about that?

Sonya Christian

Well, first of all, I really love the question because that's kind of a national discussion in terms of people having confidence in the degrees that we have.

So there are three thoughts I wanna share, so hang with me for this one, okay.

Sonya Christian

So the first one, I think the degree has always been respected as a certifying tool. If a university, you know, produces a degree, people say, oh, yeah, this person should have skills and abilities that are coming.

Now what the degree encapsulates, I think is shifting. In the age of AI, what the California community colleges are talking about is a phrase called AI readiness. And we have a specific description for AI readiness.

It has two parts. Okay. It's a plus.

The first part is technical skills, AI technical skills. You're working with an LLM, which, you mentioned our students are already doing, but now we are moving from the age of what predictive AI to generative AI, and we're already in the age of agentic AI. So those the technical skills need to keep evolving.

And the second part of that plus for AI readiness is really the ability AI learning agility. So students, because things are shifting so fast, we need to create the ability to be able to navigate the unknown and to be able to learn how to learn. So that is being built into the evolving AI literacy curriculum that we have.

So the degree needs to guarantee to the market that our students from our universities and colleges have AI ready students.

The second part to this conversation is this entire clamor about skills based. You know, forget about gen ed. It needs to be skills based, skills driven, and there's this rush I feel for everything to be skill centric.

And if you had asked me this question as a budding president fifteen years ago, I was in the race telling, you know, all my speeches, it was like, you know, we had to be skills, you know, skills ready.

Now I'm shifting the conversation in the age of AI by saying it's skills plus general education. Today, I feel general education is absolutely valuable.

There was a New York Times article that talked about, you know, the new aristocracy that can deepen, you know, the power structures. The more powerful could be the technologists who not only have the billions of dollars but also the political will.

And then there's the rest of us who are the workers who don't create the conditions for the work.

So this idea of, you know, your history, global education to me needs to be front and center in how we continue to confer the degree and that's going to be valuable.

So thank you very much.

Michael Horn

Well, I'm glad you gave the full answer there. It was good, worth sticking through the full thing.

Pradeep, the question I want to come out of that to you with is one thing I'm curious about then in this conversation around the value of the degree is it strikes me that there's a lot of worry, if you will, about jobs disappearing, but there's actually a shortage of from my perspective of deeply trained people for both entry level and senior jobs and so forth.

And in many cases, entry level jobs are getting what I'll call seniorized. That is they're expecting a higher level of skill and performance on the job.

And so I'm curious…

You know, one approach to this says, ‘Okay. well, if we have this shortage, let's do more short form training. Let's dig in on boot camps, twelve-week programs, and the like.’

And another says, ‘We actually need more general education, more multi-year degree-based education and so forth. This can't be done in 12 weeks.’

How do you think about that? What's your take?

Pradeep Khosla

A couple of things. One is whenever there's a disruption happening because of technology, the transition from the state before to a steady state after is not a twelve month transition. It's a multi-year transition, and it somehow happens in the normal course of things.

It happened when the PC became rampant, from no computing to computing.

It happened when the World Web was becoming, when email showed up.

And I can go down the list, right?

So AI is no different.

There is a disruption happening. There's no doubt about it.

Is it going to lead to massive job losses? It might happen in the transient, in the steady state. I think it's going to be a wealth generator.

If you look at this economy, this country, we have generated trillions of dollars of wealth because of technology, not because jobs were lost. I mean our employment today is higher than our employment in 1980 when we had not much technology. See what I'm saying?

So every time the disruption happened, we added more jobs. We added more jobs. Different jobs are created. Now I would not have imagined that influencer is a job, but right now if you're a podcaster, you make good money. Seriously, right?

Jeff Selingo

Not on this podcast.

Pradeep Khosla

Not on this one. What I'm saying is, so I think we need to keep an open mind.

I want to go back. So the skill in my mind, I don't look at skills as the repetitive skills of in like assembling motor cars. Skill number five and number four, we need to think about what it means. Even in the context of SAT, you getting 699 is no better than me getting 698 because you see what I'm saying? These numbers don't mean what the numbers indicate.

So I think we need to have a more deeper conversation about what college readiness means and who should be going to college and why.

Jeff Selingo

By the way, that theory is very similar to college rankings, The tenth best is not necessarily that much better than the fiftieth.

Pradeep Khosla

Exactly, and what people don't understand is higher education is very stratified. Public institutions are different than private. Privates amongst themselves are very different. There's the liberal arts privates. There's a more focused, I can go down.

So I think we have a completely distorted notion of what higher ed is. But higher ed is a part of this country. It is needed to keep on building a technologically driven economy and you cannot do it without higher ed.

Jeff Selingo

So James, let's stay with assessments for a moment because there's been a lot of press over the last couple of years about how faculty are kind of moving in different directions right now, oral exams and blue books and portfolios, right, to try to figure out how to assess students.

In the shared governance model, there's a lot of discussions though between technologists and what you're able to provide to your faculty.

And we talked earlier, faculty have the IP of their classroom.

So how is this going to work out do you think? are the faculty going to say well we're not going to use this technology to assess students because we think the best assessment is a blue book or an oral exam?

Or do you think the technologists are going to say, but there's a better way to really know whether students know X or Y?

Like, how is this happening, whether it's on your campus or where do you see it happening widely?

James Frazee

Yeah. I think, you know, again, we have to trust the faculty to know what's best for them and their students, and they're responsible for making those calls.

Like Sonya, we have faculty fellowships. We have a long history of providing fellowships to faculty.

We've also, back in 2023 for the first time ever, we provided student fellowships, AI student fellowships, and bringing those students into the conversation as it relates to this moment is really important.

It gets to, again, having conversations about what we know about how people learn, the importance of more-frequent low stakes assessments, the importance of kind of retrieval practice, distributed practice, all of the things we know from learning science that helps people learn.

And, I think that again, faculty need lots of examples of what that looks like in practice.

I mean, the possibilities today are enormous. You heard about NotebookLM this morning, that debate feature where you're getting people exposed to different perspectives on the content that they've curated to create their own notebook. And then you can … one of the audio overview features is you can have the two hosts debate the content. Yes. Talk about critical thinking and getting exposed to a broad perspective. That's really powerful.

I was saying at dinner last night, I mean, the potential to have a simulated conversation with a historical figure about a topic, that's never been possible before. And thinking about, ‘Okay, well, if you were gonna make that part of an assignment, what does that look like?’ And maybe as the student, you're being asked to assess what are maybe some of the inherent biases in what has been used to train the creation of that historical figure.

So I think you can get students to think much more critically, but the only way faculty are going to fully embrace this is if they hear about it from other faculty.

And so, I think anything that we can do to, again, celebrate those people, hold them up as exemplars is really gonna serve us well.

To the point about the value of the degree, I mean, I think there's never been a better time for the humanists than now. I think humanities is more important than ever.

People talk about durable skills. I prefer that term over soft skills. But when we talk with our industry partners, I'm part of a California State University Workforce Acceleration Board. Google's a member of that. We have a lot of industry representatives. I won't mention all the other names of the technology outfits, but they're all there around the table.

And the one thing that we're hearing from those folks when we're asking them, what are you looking for, Google, in employees? It's people who can work across silos. And people who can translate easily between these different areas of an organization.

And that's a learned skill. I think that we have to provide opportunities for students to really build those muscles and to foster the relationships.

I mean, about the value of the degree. Part of the value of the degree is just the people you meet. Right? And the doors that might open as a result of those relationships.

And if you don't have assignments in class that require you to be interacting with one another, you can't just take for granted that that's gonna happen because all you see when you're walking around campus is this is where the students' attention is.

They're not out there making friends like maybe they were …

Jeff Selingo

Well, I've always said that college may be the last analog experience that we have in life, so we have to try to keep that.

So as we get ready to wrap up here and we're gonna be turning it to the audience, we'll have time for a few questions.

Pradeep and Sonya, I wanna talk about governance because we've heard multiple… you know, have an opinion. Faculty have an opinion, staff have an opinion, alumni, everybody, students, all these places have opinions.

How the traditional ways higher education is governed, campuses are governed is slow moving and sharing.

Is that a good recipe for AI?

Pradeep Khosla

So I don't know Jeff if you are completely correct.

Jeff Selingo

Okay. It's not slow moving?

Pradeep Khosla

No. I think higher education has multiple modes of movement. It's slow moving and deliberative in the context of you want to change a curriculum, you want to change a program.

It is also fast moving in the context of administrators like me at my level. I can make choices that can be quick choices that are made. Like the choice I made to invest in AI. It was an instantaneous choice.

But after that you've got to give time to let the process mature, and that's a slow moving part of it.

So I think higher education is a combination of both fast moving and slow moving processes, and most of the people typically see the slow moving process, and the fast moving process is kind of hidden from them.

Jeff Selingo

So is our governance system working?

Sonya Christian

Well, I'll answer the question at two levels from a system office perspective.

One level, you know, our higher education institutions are microcosms of society at large.

So for example, the policies that our governor of California establishes in the legislature at the macro level is going to impact us, you know, the relationships with our industry partners in the state of California.

So what we do in the chancellor's office from a governance standpoint is make sure that we are kept updated and we keep conversations going, active conversations with the governor's office, with the legislature, so that they're thinking about policies that would then impact us in the direction we want to.

At the institutional level, we already have structures in place. We have consultation councils. I went ahead and created a special AI council to be able to fast track some of the governance issues.

And now the two conversations that I find exciting are, you know, there is an idea about the different levels. If you're using AI, if an administrative assistant is using AI to take minutes, there's not much consequence like minutes of a meeting. But if you're looking at a policy of using AI for financial aid, now the risk is higher. So how do we adjust based on the risk factor?

And the last one I heard actually a few weeks ago is there some consideration of governance for our agents, our agentic AI structure. How do we, you know, really construct the governance for that?

So it's various different levels.

Michael Horn

Alright. Last question from the two of us, and then we'll, open it up for a couple questions, hopefully, from all of you.

Lightning round, you get to fill in the blank with, say, one line, which is: Five years from now the degree is still worth it because faculty do blank that AI can't.

Pradeep, you wanna go first?

Pradeep Khosla

The degree is still worth it because faculty will adapt to use AI.

James Frazee

The faculty will create activities and assignments that require human judgment.

Sonya Christian

Because faculty are developing people, the human, and not products.

Jeff Selingo

Great. This has been terrific.

We're now going to take a quick break. Please join me. We have time for about four minutes.

Jeff Selingo

We have time for like two or three questions if that.

So there is a microphone that's gonna come around and we have a …

Michael Horn

And make sure your statement ends with a question mark. Yeah.

And if you could tell us who you are too, that would be great. Thank you.

Audience Member

I'm Achar from San Jose State University.

So do you blame the AI developers so we don't name all of them?

They ran so fast before schools and they gave you the tool and they are only working on developing features instead of bringing a good package with a lot of scenarios and teaching methodologies that will make selling the tools easier, more efficient, and not just throw the problem of privacy and cost and all that in your hand.

Do you think if they did better work in packaging solutions and if you guys sent faculty to them to talk and customize what you're looking for, would it make it easier to adopt?

Jeff Selingo

So in other words, was industry, did industry move faster than education?

Pradeep Khosla

So I think it is not right to blame developers of science and technology for developing something. Because at some point we as humans, when we accept the technology, we have a responsibility to figure out how to assimilate that into our everyday working, into the processes, into the society, right?

So I think it's easy to say, ‘Oh, if you had not done this, I would have been a better person.’ I'm not a believer.

I think Silicon Valley, the power of this country, if you look at the wealth generated by the country which has helped alleviate a lot of poverty in the world, is technology. And by controlling the pace of development and deployment of technology is not a good way to … you create censorship that I'm not a big believer in.

Jeff Selingo

Great.

James Frazee

Yeah, I agree. I'm not a big fan of stifling innovation.

And I think to the governance question, I think the deliberation that is perceived as at a glacial pace is not a bug, it's a feature. And sometimes you really have to sit with something for a period of time and mull over it and take different positions on it.

And I think that's one of the powerful features of higher education that I would suggest we don't throw out.

Michael Horn

Let's do one more.

Jeff Selingo

Let's do one more question right here. Oh over there. Sorry over there. Thank you.

Audience Member

Hannah here. I had a question.

You spoke a lot from the faculty perspective, but higher ed is of course more than just faculty. It also involves non faculty staff personnel. I'm curious how you would incorporate AI literacy, because I believe that's more of a career readiness skill in addition to technological skills.

What's your insight and word around how faculty and non faculty can better collaborate within the classroom because that's something, one of our pain points as a career scene, students don't come find us until it's too late, right.

So how do you, tips or tricks around how faculty and non faculty can kind of target and support students sooner than later?

Jeff Selingo

Anyone can take that. We don't have to all three of you.

Sonya Christian

I'm happy to jump in. Our employee group across California Community Colleges, our largest population employees are our classified professionals and they are the ones that engage with the students.

So when you look at the resources we have put into professional development for our classified professionals, it's significant. It's absolutely critical because that's the first point of contact for students.

The other dimension to it is California colleges and I think the universities as well, we're unionized. And so entering into agreements to make sure that we are protecting jobs, but really looking at skills that have to evolve.

So really counting on that evolution of skills with AI and technology, whereas giving the security of jobs, keeping up with that evolution.

So those have been some strategies we have adopted to motivate our employees to really work in engaging with AI.

Jeff Selingo

Terrific. We're going to leave it there. Michael, any closing, reactions?

Michael Horn

Sure. I'll just two reflections that come out of the conversation for me, Jeff.

One, there's this sort of notion of we're building humans, and we are all humans, at the same time.

And I guess the headline I'll have is ‘adults wanted’ or maybe it's created because we want room for the trust, room for the conversation, and what can appear like incoherence from the outside is actually friction that is a feature, not a bug. Is one thing I'm hearing.

And then the second thing I heard is this tight loose formulation you spoke about, Sonya.

And it reminded me a lot of the conversation. Coach K, Mike Krzyzewski, a former coach at Duke, used to say he doesn't have rules for his team. He had principles for them to follow.

It's a good way to think about governance in this year as well.

Jeff, what about you?

Jeff Selingo

I had three quick ones.

Radical transparency helps trust.

Second is never underestimate the creativity of faculty. Again, trusting them.

And then finally, we are in a disruptive state right now, and we kind of have to accept that. There's a steady state disruption, and there will be a steady state at some point.

And I think that higher ed has been in a disruptive period essentially since the pandemic, and I think people are looking for that steady state. Hopefully, it will come pretty soon.

So again, thank you to Google for Education for having us today.

Thank you to our panelists here.

And thank you. Thank you all of you for joining us on this live recording of Future U.

Pradeep Khosla

And the hosts. Thank you.

Jeff Selingo

Thank you. Yeah. Thank you.