Entrepreneurship | The AI-enabled entrepreneurial university: How GenAI could transform venture creation

If we don't use AI to improve how we educate, people will get their education elsewhere and then we won't have universities anymore. Hello and welcome to the Entrepreneurship Podcast series. In this episode, Todd Davy, associate partner at UIN, sits down with Scott Shane, professor of entrepreneurial studies in economics at Case Western Reserve University. They explore how generative AI could reshape university research, teaching venture creation, governance and institutional productivity at scale. Okay, welcome everybody. Today we will be speaking with Scott Shane and what we'll be talking about is the AI enabled entrepreneurial university, which is a topic of conversation that Scott created together with AI. So I think that's going to be an interesting aspect to this conversation. And what we're going to be talking about is how Gen AI could transform venture creation and also the various aspects within the university. My name is Todd Davy. I'm an associate professor of entrepreneurship at the Institute of Intellicon Business School in France and also associate partner at UIN. Without any further ado, I'm going to introduce Scott. Scott Shane is the professor of entrepreneurial studies as well as professor of economics at the Weatherhead School of Management at the Case Western Reserve University. He is a managing director also of venture capital fund investing in early stage companies. He has published many scholarly articles according to his website, so over 94 now. But there's two in particular that have relevance to us, which are the illusions of entrepreneurship and academic entrepreneurship university spin-offs and wealth creation. Thank you. Thanks for having me. Thank you very much for making yourself available for today's discussion. Really looking forward to getting into some topics because I guess that's something that I've always admire about your work and your research is that you've been really dedicated to converting your research into meaningful policy and practical insights. So I guess I want to start the conversation there. I want to understand a little bit more about what has proven you to bring your research into practical use into a setting where many academics might not. It's an interesting question and it's very salient for me right now because I'm about to get back the copy edited manuscript of a book called accidental venture capitalist, which is a book that MIT Press is going to bring out later this year where I talk about my experiences starting a venture capital fund and using it to draw on all the venture capital research I did. And I guess I think about the answer to this question in two ways. The first way is social scientists are in certain respects outliers in doing this. If you took most biological scientists at universities or engineering researchers at universities or computer scientists or others, they wouldn't think these things were separate. They would think that they're the same right people invent novel semiconductor designs and because that's their research and then industry as it says we can have AI because this chip that this person that invented like allows for inference better or something like that. And so that's normal. So we may be the outliers and honestly I think it's very important and I do it because I think it's important and because I want to give it the test of reality. I want that ecological test of whether there's validity to the things that I say. And I think the problem is parts of academia kind of aren't relevant and we allow it to stay when we probably shouldn't. Okay that's a interesting comment because I'm going to pick that up a little bit because I also wanted to get a little bit more. So that's why you do what you do but what do you think or if you had to elaborate on what are the role of the modern academic in society? What do you what how would you explain that? So academia has two functions one of which exists through the whole university and is the older one which is education right and education means education of our students and creation of new knowledge that diffuses out to everyone and educates people makes people know more than they used to know before people taught them to do something new. In the more modern era and I guess I would define this as when Bismarck decided Germany needed or what was to become Germany needed the infrastructure of universities for producing useful things and that was to solve problems like a problem in agriculture that people didn't know the answer to that would help improve agricultural yields or a problem in industry of making things out of novel materials or something like that because the industry needed to make things out of novel materials and this goes on into a more modern space. The part that I say about this is everybody in the university educates including parts that may not have that kind of useful productive direct effect on economic activity around them. The humanities are further removed right there's probably lots of good reasons we want to understand what made Shakespeare a great writer and learning to write well is probably something useful. I don't know how practical that would be in the sense that metallurgy would be really super practical to I don't know making the rails that railroad is run on right and that's the case. I think both of these are important and I think one of the functions of the modern university is to do that non educational piece because technical advance is complicated and you can't do it without some of the best researchers and thinkers that society has. I think you provided a really nice segue there when you meet in the modern university because I think there's a great segue into your contribution that I want to reference because it's part of our future university's Thought Book initiatives that we have been running since 2018 now I think we're into our sixth edition and you provided a really interesting article because you also approached it about the topic of AI and how AI enabled entrepreneurship but also entrepreneurial universities but you also used an AI tool to do that so maybe just to start off why did you decide to use an AI tool to do that as that particular topic and what I want to preface that with is by saying I have constructed your questions in following your lead on this I've constructed the questions using AI as well. Yeah for sure so I'll give you I'll give you the whole context so first of all this contribution to the Thought Book is not the first contribution to one of the Thought books that I ever done the first one I did was in the pre generative AI and by that I mean pre to the real public because other people were using things many years before but we didn't know they were using them if you were just regular person but so the first part was that and I did this by writing with a clinical professor at my university and I realized that when I got asked to write for this Thought Book that I didn't need that person to do this because there was a substitute and the substitute was AI I had already been using AI for lots of things when I did this I'd be using AI to help me grade open ended student assignments and had used it to help me write papers write books etc. Now what's interesting about this is I did this with chat GPT 3 points something version I don't remember the exact number but within the three numbers which was only like maybe 18 months ago max when I started maybe even a little less and we are many versions better than that now and the big difference is it was really useful for doing the particular project that I was working on and it was for this project and what's happened since then is it got better for other kinds of things that are more complicated than writing a short essay. I think that's also a really nice step into part of your article because it describes their example that Gen AI will fundamentally reshape universities by 2035 and that was one of the major themes so I just wanted you to explain what you see those changes to be and where is that going to take us, whether it's a higher education sector or just generally in society. Yeah so I'm going to make two comments the first one is when I wrote that and this was like a while ago I think I was shortsighted on the piece of change so if I had to say it now I wouldn't put not the only one on the way. 35 I would say we're talking for sure 2030 and maybe we're getting into the late 20s right it's very fast. The second part is it is affecting all aspects of the university and I define those aspects in four categories. I define the scholarly research discovery of new knowledge and testing hypotheses that teaching part of informing students of information and testing them on that activity. The service component of the university like the collective governance of keeping the institution operating there and then the final one is the thing when we talked about the modern university of contribution outside of it's all to the broader economy and society right. All of them in effect. and it affects it in different ways, right? So, and the rate of changes of differential peace and differential reaction across the board and all of them generate some kind of transformative effect. Well, I think what I wanted to do was I originally had a plan, and as I said, I recorded the questions with AOA, but I think now that you've provided that structure up, I thought it might be worthwhile just going through each one individually just to understand a little bit more about where you think the changes are going to occur. And I guess there's going to be some positives that are going to come out of it, and there's going to be some other, I don't know if we call it negatives, but the consequences. So, let's start with scholarly research and discovery. Can you perhaps explain what as the role that AI might have and how it will change things? Because your contribution really elaborates on that really clearly. Yeah, so I think the best way to do this is explain things that I am doing with the AI, which I think it does as an effect. So, just to be very specific, AI is much more logical than humans. And so, if what you want to do is write a logical argument that says, "A leads to see the gets to this hypothesis," and then there's this data that's going to then support or falsify that hypothesis. AI picks up the contradictions in the exposition really well and will draft it better than humans and will keep you from overstating causality and all things like that. And so, this is number one. I think we're seeing this in all this academic work, and that is improving both the pace and quality of research work. I mean, it's like putting a researcher on steroids. So, I would tell you that my prior peak of research was when I was in my late 30s and early 40s around 2000, and I am now more productive than I was at that pace, and it isn't because I am spending more time than when I was untenured or cared more now. I'm more being the senior scholar. It just makes you that much more productive. And everything that it substitutes for, it runs your regression analysis, formats, tables, does all of that better. It also is increasingly, as we move forward, allowing you to test and rule out failed paths. If I run this experiment and it works perfectly, how big a magnitude of a finding am I really going to have? And if I'd find out, ah, that's going to be small, maybe the opportunity cost is a little high for picking that path and going to do this other one. Right? And so, those kinds of things are in research, and this is everybody, and we see this in the data. From before chat GPT's big 2.0 release to now, across fields, scholarly paper submissions to academic journals are up for fold. That's enormous magnitude increase of research. And so I'll just stop on that one and say, that's the example, just literally using it to produce the knowledge in ways that are far more efficient than the way we did before. And what I found interesting when you were describing it in the article was it wasn't just about writing. It was actually about the execution of research and the speed at which you could move if AI could support that process. And I thought that was really insightful. I want to pick up that thread a little bit later as well, but just in respect to education, perhaps you could just touch on that, where do you think the role of AI and education will be? Yes, so there's multiple things on education. So first of all, there's the whole idea that AI allows you to scale customization in ways that human people don't really want to do things more than they have to. No professor really would rather like be preparing 25 different versions of something that they're going to teach tomorrow for their students instead of like having dinner with their family. It's not actually a choice any of us would make. Absolutely. But it is definitely the case that if one student doesn't really understand the thing you did the week before that this is building on, you can't start with the assumption they understand it. And this other student can and you just can create these customized versions by writing a prompt in cloud or chat GPJ. And this is and nor the customization is enormous increase which will dramatically affect learning because if you tell people I'm teaching you to play tennis and you have to actually have no idea to grip the racket because you didn't learn that when we did the first tennis lesson, you're not me talking about how you're going to hit the back hand when you still don't have the grip is really a very bad idea in the tennis lesson. What would be better ideas me to like give you the grip thing while not making the other student who already is basically on the junior varsity team. Learn how to improve their backhand to get it really good, right? Like that's that customization. The second thing is we don't do well on certain kinds of feedback because we're humans. In class when we have a discussion, the ability to evaluate simultaneously some students answer and give them feedback is far less than my ability to read their essay and give them feedback. But if I tape a record that session, AI will run through that transcript and get exactly that feedback on the verbal session which I didn't do well. So humans are relatively better at evaluating reading written documents than they are in evaluating statements. Let's put it that way. And there's also we find it easier to be unbiased by looking at an essay and not paying attention to whose name is at the top. But I cannot listen to a person in class and not recognize that person is that person that they are and they they're a human and I'm going to react to all their human attributes. Right? So that's it. This part. The other part is where we have a hard time with other things about understanding. When people are unclear in their communication like a student asks the question, it's hard because we don't understand it as people and AI will be able to figure out what they said and often give them a better answer than the human would. And part of being a professor isn't just spewing out information but answering questions. And that is there. The final thing is on the evaluation part. Human beings are actually quite terrible at evaluation. I can't be mean when students cry in my office. What that means is people who cry get better grades than they probably should. Also, when I want to leave and go home and somebody's in my office hours and they come up with the end's objection to something I give in so I can go home and AI doesn't do that. And so many things on the evaluation side are better. Well, I think there's some people that would probably say you're just being human. And there's a human part of this interaction that we'd like to keep. But I want to pick up the third part which in your article I thought was really strong. It came across for me at least very strong that one of the first areas that the university is going to be transformed is in respect to entrepreneurship and how this will lead the transformation. And you talked about always on venture studios. And perhaps you could explain that a little bit more about what venture studio model is, what that means and then how could that practically work? Yeah, so okay, universities are places where people come up with discoveries. And sometimes those discoveries aren't really important in a scientific sense. They're important in a commercial sense, right? I figured out a better way, like for example, I'll give you an example of a company that I know that came out of Hays Weston Reserve University where these two who were PhD students at the time came up with software to control robotic arms so that you could do better welding. And welding is a complex problem and the software could do that. It's not like this is a Nobel Prize winning pure research. It's that man, the average age of a welder in the US is 52 or something like that. And we're not replacing welders. People don't want to become welders and we got to come up with robotics to solve this. And the commercial sense. So the thing is that they go to places like a maker space on campus and start a business and talk to investors and all of this. And so why does AI help on that? It really helps in a couple of dimensions. One dimension is that is an exact example of customization. By definition, you don't want to start a business that is identical to somebody else's business because you don't get that needed. And so if we try to teach them to do something that's identical, like we're going to take 10 people we're going to have them make exactly the same business. We're actually doing them a disservice and the idea is the AI has the patience to do that. University can't afford 10 different mentors. Each one working individually with one startup is just the economics don't work. The second part of it is most things won't work. And what you want to do is effectively do an in-silico test of every idea before you do it. Instead of making something and going out and traipsing to people and saying, "Do you want to buy this thing?" you should figure out if you could actually make it, right? With this work hypothetically, because if I can't make it for less than the, if I can't come up with a solution that says what the cost people will pay, the price that people will pay for, I can't produce it for less than that, nobody's going to buy it. I can't give the economics. It's not going to work with just these ideas out. And AI is really good for the umpteen iterations to figuring things out before you launch them. And those two things alone, I think are the Venture Studio example. And then the only other part about the Venture Studio is honestly, this is where the big money part would be because people who start really successful businesses become wealthy, they would give back donations. Companies, the universities have figured out a long time ago that they should take equity in some of these things or charge licensing fees for the rights back to the IP. And there's a money in that stuff. And like anything else, if you don't make money, you can't. Your institution will go out of business. Yeah, no, I think there's some real insights into the customization and the iteration part because I think that you touched on that and the fact that if you have an idea, it has to be unique, of course, but be able to customize the approach that you're taking and then also to make that work faster. I can see that clearly. The question I've got though is why are universities particularly well placed to play this role? Why is this happening at universities? Why isn't this happening elsewhere? That's just the general notion of roles in the stage so that you can't really get the private sector to expend resources early on on things because chances are any given activity will generate a return. So that's reason number one that basically universities have another mission. So it just changes the economics a little bit that they can do it at a lower return than the private sector would be willing to do. The second thing is that they're a safer actor to taking government resources. There's all kinds of favoritism problems when you say, "Let me go pick some company to get the government's money." That's very much harder than saying, "I'm going to give it to these public institutions that are designed for education and they're going to dole it out to the individuals and make this stuff happen." It just that works better. And then the final one is just a practical one. The 10-corers and the maybes tend to be people who are being educated because if you're already working on something commercial, you already have a path that probably is less likely to lead you there. A student is more likely to discover something that's going to overturn IBM than the IBM employees and discover something that will overturn IBM. If we go down that path, then what do you see is the key benefits and for whom? Maybe also just think about the other side, but what could be the potential downside of going in this direction? I believe that technological progress of any kind on average provides more benefits than its cost because people on average wouldn't do it and wouldn't allow it to continue if it was worse. We must believe things like the internal combustion engine is better than horses and buggies even though it causes some kind of problems in society because we would have stopped it if we didn't think that. I think it's true also for AI, so I think it's net beneficial and the problems and we'll just have to come up with guardrails. I'm not the expert on the very specific guardrails that we need to minimize the problems and maximize the benefits, but I'm just going to say on balance, it's going to be better for society and then get into why would universities do it? I think it's because it's also going to be better for them in their particular role and many of the reasons I was describing earlier. And someone argue that universities need to take this step as in need to modernize, need to play a broader role in society, would you agree with that? Universities are rarely at the forefront institutions on any of the jobs they do. So I would tell you that they are the lagers on any new educational mechanism that's used, right? If they weren't pulled into it, we would still be teaching the way Plato taught Aristotle. We would still be stuck with a stone tablet there, but we got pulled into it because outside the university, people were doing education different ways and people were like maybe we shouldn't go to universities, then in the university got sucked in, right? It's basically if we don't use AI to improve how we educate, people will get their education elsewhere and then we won't have universities anymore. I want to just explore one topic a little bit longer and then we'll come back and finish up and tie it up with the universities. But if AI dramatically reduces the time and costs required to validate ideas, build prototypes, what should founders and entrepreneurial teams be good at in this scenario? And how should universities be supporting them? So the question is starting a business being an entrepreneur is all about the world has some problem that no one else solved and I can come up with a better way to solve that. The AI can help you both figure out what the real problems are from the mess of everything else by dramatically increasing the ability to gather information and reason about that information and then come up with the solution by gathering and reasoning about that solution. And to me that's the heart of it for the founders is like it's just, it's like a steroid for getting rid of some problem that doctors assigned steroids to. That's effectively what it is. Or in the VC world it's always like the VCs are the money is just putting gasoline on the fire. It's the same analogy. Okay, I just want to now because we've gone to our time limit, so we've got probably three or four minutes left. I just wanted to explore a little bit more about if we want universities to play this sort of a role. You as you mentioned, not knowing for being moving fast and the bureaucratic nature of universities would indicate to me that might be a role that they would struggle with. What is the single biggest challenge do you think that the university's face in actually taking on that sort of a role in adopting AI and moving in that direction? That they don't want it like it's as simple as that that they have to come to accept in every area that it's here to stay and they have to figure out the best way to deal with it. Not fight it. I'm reminded of there's this meme that goes around that shows pictures of professors in the 1970s standing around, and they're kicking and protesting the use of calculators. And that meme is the example. People are saying we shouldn't have AI, AI is evil in the universities. It's bad. And if all that energy is going to get wasted fighting it and we're not going to figure out the right way to use it, that is the current. The stream is flowing fast that direction, learn to ride those rapids. Don't try to go upstream. Well, I mentioned today that I got AI to write me some questions to be quite honest. I banned and many of them really didn't quite fit. But this one in particular I think is quite interesting because it's basically asking you, are you ultimately optimistic or pessimistic about the future of universities in the age of Gen AI and why? I am optimistic because at the micro level, Gen AI has made me more productive at absolutely everything that I do regardless of whether it's work or not. And that productivity, I had you have to be optimistic because it's either I can do more of what I am doing and therefore get rewarded for that. Or I can say, I have all this free time I'd rather do something else with my time and that's got to be a good thing. Like I just I don't know how you can be pessimistic about that. I'm just thinking some of my colleagues and I think probably 50% would fit in one category and 50% would fit in the other one. But I'm just going to record up now. I'll be the final question. And I guess it's just in respect to university leaders listening today. What would be you would suggest somewhere between one to three important actions that you think they should take as a result of what we've discussed today. Get out of the way. Let the system naturally emerge. Get out of the way is number one. Number two is to the extent you want to do anything. It's figure out where the problems start to emerge and put up the guard rails. And those are the only two that I can see. But the number one thing that I see happening is standing in the way it's saying you can't do X or Y in teaching or search till we figure out our policy towards it. That is the get out of the way and don't do that. Professor Schatz, Scott Shane, has been an absolute pleasure. As always with you, we have been talking about the AI enabled entrepreneurial university. But again, I just wanted to take the opportunity to thank you profusely Scott because I think some of your thoughts are really going to help people who are grappling with these universities. just thought the contribution that you made to the Feature University Thought Book was really interesting. to stretch our thinking. So thank you very much for that. Oh thanks for having me, I appreciate it. And thank you to everyone who attended and those that are listening online and are podcasted, you can check it out. Thank you very much to everybody. Thank you, James, for Scott. Thank you for listening to today's discussion. Follow UIN on LinkedIn and if you're enjoying our podcasts, make sure to subscribe, rate, and review in your podcast platform of choice to help other people find this content too. [Music]