‘20 PhDs’ In the Time of One: How AI Is Changing College (YouTube Video Transcript)

(00:00:00) Your YouTube transcript will appear here (00:00:00) We decided to change everything. Change (00:00:01) culture, change design, change (00:00:03) intellectual structure, use technology, (00:00:05) move in new directions, and then we (00:00:07) began measuring learning outcomes. (00:00:09) >> For all the talk about changes in higher (00:00:11) ed, including just this [music] week (00:00:13) when the University of Virginia became (00:00:15) the first public college to agree to (00:00:16) Trump administration oversight, the (00:00:18) biggest cause for change, AI, doesn't (00:00:21) come up all that often in the political (00:00:23) debate. For nearly a quarter century, (00:00:26) Michael Crowe has led the university (00:00:28) that is by some measures the largest in (00:00:30) the United States. Few people have (00:00:32) thought as much about American higher (00:00:34) education while also having the ability (00:00:36) to influence it on such a large scale. (00:00:39) But when we first talked with him a (00:00:40) little over a year ago, we didn't cover (00:00:42) artificial intelligence. A lot has (00:00:45) changed since then. [music] AI will (00:00:47) leave a lot of white collar people (00:00:48) behind. (00:00:49) >> Another uncomfortable [music] truth (00:00:51) linked to AI. (00:00:52) >> AI. AI. AI. (00:00:54) >> AI. (00:00:54) >> It's not a bubble. (00:00:55) >> Obviously, the most disruptive (00:00:57) technology in the history of mankind. (00:01:00) >> Egalitarian access to knowledge is at (00:01:01) the highest level in the history of our (00:01:03) species. What we have is a walking, (00:01:04) talking, reference library on any (00:01:06) subject, and we never had anything like (00:01:08) that in our society before. (00:01:10) >> We wanted to know how artificial (00:01:11) intelligence is changing the American (00:01:14) college experience, what it means for (00:01:16) the teachers and students who have now (00:01:18) made it a regular part of their lives. (00:01:21) But we also asked Crowe about the (00:01:22) outcomes for recent graduates and how (00:01:25) his school is preparing students for an (00:01:27) economy that is moving very fast. (00:01:30) >> What does that do to teaching? I mean, (00:01:32) back in the olden days, we wrote essays. (00:01:34) We wrote blue books back when I was (00:01:36) there. How do you do things like essays (00:01:38) and evaluations? (00:01:39) >> Well, I think what has to happen, and (00:01:40) we've experienced this at ASU with our (00:01:42) 6,000 faculty members, several thousand (00:01:44) of which are already AI trained, is you (00:01:46) have to up the game. Perhaps we were (00:01:48) learning too slowly, too incrementally, (00:01:50) too much in a regimented or industrial (00:01:51) way. With the AI tools that are (00:01:53) available now, you can up the game, uh, (00:01:55) enhance the question complexity, enhance (00:01:58) the answer complexity, expect more of (00:01:59) the students. We had somebody uh, give (00:02:02) their test out of the business school to (00:02:03) an AI system and get everything right (00:02:05) instantly. Well, then the test is too (00:02:06) easy. The test is too simple. So, you (00:02:08) need it's basically a way in our view to (00:02:10) accelerate learning, to broaden (00:02:11) learning, and to speed learning. So, you (00:02:13) have to look at it as a new way to (00:02:14) basically make the game more intensive. (00:02:16) The model's always changing. So Plato, (00:02:18) you know, was against the written word. (00:02:20) He thought everything should be uh (00:02:21) thought through uh verbally and (00:02:23) communicated verbally. There were (00:02:25) unbelievable forces against the (00:02:27) development of the printed book. And so (00:02:29) uh the internet in its development, the (00:02:31) web and its development all had people (00:02:33) that were against it. And so AI changes (00:02:35) the model in the sense that it speeds it (00:02:37) up and intensifies it. It personalized (00:02:39) the learner's experience, but it doesn't (00:02:41) teach those core things. There's no (00:02:43) values being taught. There's no values (00:02:45) being experienced. There's no lived (00:02:46) experiences being built. So what we (00:02:49) really have here now is we just have (00:02:50) this massive hypers speed calculator (00:02:53) capable of going to all of the digitized (00:02:55) information. You asking a question about (00:02:57) that information and getting the most (00:02:59) probable answer. It's all about the (00:03:01) questions that you're asking. It's not (00:03:03) about the answers. It's about the (00:03:04) questions. And that's what people need (00:03:05) to really figure out. (00:03:06) >> Does it change the notion of cheating? (00:03:08) >> I'll bet humans have cheated for quite a (00:03:11) while. It does change the nature of what (00:03:13) is your work. Now, if you're answering a (00:03:15) complex question and you're using a (00:03:17) reference library and an AI system to (00:03:19) answer that question, that seems (00:03:20) legitimate. If you're if you're using it (00:03:22) to produce your analytical response (00:03:24) that's supposed to be demonstrative of (00:03:26) your ability, well, then you're you're (00:03:28) you're cheating. Uh now, you have to (00:03:30) then build a system which recognizes the (00:03:32) ability to gain access to these tools. (00:03:34) Now, sometimes it's just going to be you (00:03:35) in there by yourself taking the test (00:03:37) because they've got to know that you (00:03:39) know how to ask the question, you know (00:03:40) how to derive the answer, that your (00:03:42) brain works in a certain way. Beyond (00:03:44) that, the AI systems are going to (00:03:45) enhance learning in every possible way (00:03:47) and that the idea of cheating will (00:03:48) change. (00:03:49) >> At this point, are there some things (00:03:51) that you can learn that AI cannot teach (00:03:54) to you? (00:03:54) >> Absolutely. I mean, uh, you know, an AI (00:03:58) system can't teach you to be innovative. (00:04:00) It can't teach you to be creative. It (00:04:02) cannot teach you grit. It cannot teach (00:04:03) you the discovery process. It cannot (00:04:06) teach you I mean, it's a machine. It's a (00:04:08) it's an advanced hypers speed (00:04:10) calculator. Uh, it can do things that (00:04:12) you can't do. It can think around (00:04:14) corners that you can't see. Uh uh but of (00:04:17) course so can a so can a dog. And so and (00:04:19) so it's it's it's a powerful analytical (00:04:22) tool to enhance our mental capabilities, (00:04:25) not to replace them. (00:04:26) >> In order to ensure that AI is a (00:04:28) springboard rather than a crutch, Crow (00:04:31) says students and teachers will have to (00:04:33) raise the bar. And one place he's (00:04:35) already seen signs of AI's ability to (00:04:38) supercharge progress is in the school's (00:04:40) research programs. It's almost (00:04:42) unbelievable. We have probably 50 (00:04:44) research groups that are using advanced (00:04:46) AI to solve unsolvable problems to (00:04:49) figure out how to process materials or (00:04:51) manage the Mississippi River in a (00:04:52) different way in terms of the flow of (00:04:54) the water and the flow of the dirt and (00:04:56) other things that go down the river. (00:04:57) We've got people doing advanced (00:04:58) chemistry now. We're using AI systems to (00:05:01) think beyond the way that we normally (00:05:03) think to create more revolutionary (00:05:05) opportunity. uh there was a a study (00:05:07) recently done by some of our faculty at (00:05:09) the speed with which you could complete (00:05:11) the work equivalent to a dissertation in (00:05:13) in uh genetics 14 days. What that means (00:05:17) then is that the PhD that normally takes (00:05:18) four or five years to set up the (00:05:20) experiments, do the experiments, do the (00:05:21) work, be evaluated. Maybe the PhD (00:05:23) student of the future will do the (00:05:25) equivalent of 20 PhDs. That will speed (00:05:27) up the cures for cancer. That will speed (00:05:29) up the analytical tools that will help (00:05:31) restore sight in human beings. that will (00:05:33) that will speed up the techniques that (00:05:35) use electromagnetic current to affect (00:05:37) people with Alzheimer's disease and (00:05:39) other neurodeenerative diseases all of (00:05:41) which are computationally limited (00:05:43) >> certainly the world has changed (00:05:44) enormously since I came out of college a (00:05:46) long time ago now but my sense is the (00:05:49) rate of change has really increased (00:05:52) maybe even geometrically how fast it's (00:05:54) changing how do you prepare graduates (00:05:56) today for a world 20 30 40 years down (00:05:58) the road that I can't even imagine (00:06:00) >> so there's there's no way to prepare (00:06:02) someone for something you don't know (00:06:03) what it will be except one thing. What (00:06:05) we call we're we're attempting to take (00:06:07) all the people that are coming to our (00:06:08) university, 120,000 degree seeeking (00:06:11) students, 700,000 other learners who are (00:06:14) just taking courses with us digitally (00:06:16) and otherwise. Can we help create you to (00:06:18) be a master learner? Can we help you to (00:06:21) be a person capable of learning (00:06:23) anything, adjusting to anything, (00:06:25) adapting to anything? It's really that (00:06:26) because we don't know what all of the (00:06:28) adaptations that will be required are. (00:06:30) We do know that you should be grounded (00:06:31) in, you know, American history and (00:06:33) economics and the role of democracy and (00:06:36) and certain subjects in math and science (00:06:38) and so forth and so on. And then after (00:06:39) that, we find a learning path for you to (00:06:41) take where you learn to learn. We don't (00:06:43) care what your major is. You can I met a (00:06:45) kid the other day was majoring in opera (00:06:47) and physics. Great. Fine. Fantastic. (00:06:49) That's how that kid learns. And so and (00:06:51) so that's what we're after. How do you (00:06:53) how do you create universal learners (00:06:55) capable of learning anything? And that's (00:06:56) the pathway. (00:06:58) >> We hear at 40,000 ft about a shifting (00:07:01) employment situation for recent college (00:07:03) graduates because of AI. Are you seeing (00:07:05) any of that in the real world? (00:07:07) >> We're not we're not seeing that in our (00:07:09) graduates. Now, the problem with people (00:07:10) talking about all college graduates, (00:07:11) there's more than 20 million people in (00:07:13) college. A couple million go to what you (00:07:15) think of as sleepaway colleges. You (00:07:17) know, they go they go to places where (00:07:18) you're, you know, you're living on (00:07:19) campus. The other 18 million go to (00:07:22) college in some other way, community (00:07:23) college, online, some kind of other (00:07:25) course, and so forth. So we're not (00:07:26) seeing any change. Uh you know we're (00:07:28) seeing the same level of anxiety. We're (00:07:30) seeing the same level of the process. (00:07:31) We're you know more than 95% of our (00:07:34) students that graduate as undergraduates (00:07:36) are employed or in graduate school (00:07:37) within the first year. Uh almost all (00:07:39) within the second year. Uh so we're (00:07:41) we're still you know seeing good ROIs. (00:07:44) But what we are seeing is students you (00:07:47) know who are quite savvy you know (00:07:49) adjusting their trend. So we're seeing a (00:07:51) slightly downward trend in computer (00:07:52) science and a slightly slightly upward (00:07:54) trend in double majoring and triple (00:07:56) majoring. Uh uh more uh people moving (00:07:59) into analytics and supply chain and all (00:08:01) kinds of other things. And so the market (00:08:03) for learning is also adjusting. (00:08:06) [cheering and applause] (00:08:08) >> We measure our success based on who we (00:08:11) include. (00:08:12) Crow hopes the size and scope of ASU (00:08:15) will help with that adjustment, allowing (00:08:17) students to react quickly and build new (00:08:20) skills for the changing world. [music] (00:08:22) And in a school made famous for opening (00:08:24) its doors rather than being exclusive, (00:08:27) he thinks the most important skills of (00:08:29) all can come from unlikely places. (00:08:32) >> We've even got ways now that we're using (00:08:34) advanced AI enhanced robots to help (00:08:36) people that aren't qualified to get into (00:08:38) a particular college to do what they (00:08:40) want to do to get them qualified. Guess (00:08:42) what? When we get them qualified, they (00:08:44) have more grit and determination than (00:08:45) anyone else who sort of walked into it (00:08:47) from high school and they outperform (00:08:48) everyone. (00:08:49) >> There's a theme going on right now that (00:08:51) maybe college has been overrated, (00:08:52) oversold. What do you say to parents? (00:08:55) >> What we have is a way for your your (00:08:57) child, your student to learn on the path (00:09:00) that's going to enhance their ability to (00:09:02) be most adaptive throughout their life. (00:09:04) So, don't worry about their major. So, (00:09:05) we get these parents that say, "Well, my (00:09:07) kid needs to major in accounting so they (00:09:09) can get a job, or they need to major in (00:09:10) anything other than political science or (00:09:12) history or English, where they'll never (00:09:14) get a job." That's actually not true. (00:09:15) Some of the hottest things that we have (00:09:17) that we're producing right now are (00:09:18) English majors that can code. And so, (00:09:20) they have a broader perspective and they (00:09:23) can code. And so, we provide free coding (00:09:25) classes to everyone in the institution. (00:09:26) We provide other ways in which you can (00:09:28) double major, triple major, take other (00:09:30) kinds of things. And so what we say to (00:09:31) parents is (00:09:33) let's find the way where your kid is (00:09:35) going to smile while learning, while (00:09:38) preparing themselves to be a master (00:09:39) learner, and you'll have to worry about (00:09:41) them less. If they take a fixed thing in (00:09:44) a fixed way, on a fixed pathway, uh they (00:09:46) could find themselves in an alley and no (00:09:48) way out. We're trying to make sure that (00:09:50) that doesn't