The value proposition that universities provide is (a) in person contact and (b) attendance confirmation. Claude would fail my finance classes leading to a CPA because I only do them in person and I take attendance (at the undergraduate level).
Kids are glued; AI is only another screen, though worthwhile.
On the K-12 level, high achievers who might most benefit from an AI tool need socialization skills. Schools often provide food, find abuse, provide community sports opportunities, and connect students with future employers. Claude isn’t doing that.
Hollis, this is a fantastic and provocative piece. The "Claude Test" is a brilliant diagnostic tool that perfectly exposes the vulnerabilities in the current higher education model. I agree, universities market themselves as bespoke centers for human transformation while often operating on an industrial model of content delivery, especially in the online, asynchronous courses that Claude could so easily ace.
If an entity with no lived experience, no capacity for growth, and no genuine understanding can pass a course, then that course is not assessing human learning. It's assessing information synthesis and pattern replication. Yes to this: a call for universities to double down on mentorship, community, and ethical development, the real "magic dust", is exactly the conversation we need to be having.
Where I would offer a counter-argument, however, is with the prescription to "jettison Gen Ed" and outsource it to AI.
I guess, the problem isn't higher ed as a concept, but its current, often impersonal delivery. By outsourcing it, we risk turning the foundational stage of higher education into a sterile, check-box exercise, as you suggest, thereby depriving students of the very chance for discovery that might shape their entire future. We would be conceding that the start of the journey doesn't need the "magic dust," when in fact, that's where it's often needed most.
Perhaps a better solution isn't to unbundle higher ed, but to re-infuse it with the human-centric values you champion. Let's use AI as a powerful tool, a "super-TA", to handle the rote aspects of these courses, freeing up faculty to run smaller, Socratic discussion groups, even for introductory material. The challenge issued by the Claude Test shouldn't be to cede this territory, but to transform it into something Claude cannot pass: a genuine exercise in human curiosity, discovery, and connection.
Is the complexity or advancement of knowledge *within* fields necessarily isometric with their teaching, though? An advanced seminar in history, about a very specific period or edge-case topic, might need less personal interaction than the 100-level courses where students are taught how to think like a historian, albeit with well-worn “content.” Unfortunately, departments often save their best teaching for advanced majors who have shown that they don’t need it, while the intro-level courses are considered low-status and menial at the supply side, and therefore regarded as such on the demand side. In other words, the personal transformation might happen at the start of learning a subject, rather than the end.
Also, to what extent is this also an assumption about *speed* specifically, and not just organizational efficiency? You can learn piano in a year, but 12 very good teachers can't get you there in a month.
Gen ed has been built with no theory of *mind* and no theory of *knowledge*—just liberal representational politics, but applied to academic disciplines instead of people.
An interesting article ,as usual! It seems there are two connected issues here.
1.Why not do Gen Ed in high school,exploiting AI ? High school in the US is not a primarily academic enterprise ,with a few notable exceptions (ie Bronx HS of Science,etc) ,but rather a rite of passage.. Most intelligent,motivated students are bored stiff and basically marking time .There simply are not enough qualified teachers to meet the demand for AP and other such advanced courses.Doing Gen Ed by AI could rectify this situation.It could also shave a year or more off "college" and,more importantly, allow students to spend their college years exploiting the real resources of universities.
2.What to do in college? With mass produced Gen Ed out of the way, students could partake of advanced,specialized classes that actually utilize faculty expertise.More importantly,students could get involved in research and other experiences that feature personal mentorship by faculty ,thus leading to substantive intellectual and personal growth.
I took this month off from Substack (mostly) to read books, so just coming to this one.
It brought to mind something I found in my reading. Walter Lippmann explaining Henri Bergson's appeal: "He is popular because he says with splendid certainty what thousands of people have been feeling vaguely."
If an AI chatbot can earn a degree, then we need to make significant changes. I think thousands feel this vaguely. I hope you saying this with splendid certainty means you your way to becoming popular.
Well, we've been talking about that tough nut for a while, right? I think it all comes down to that old Steve Martin joke that the key to comedy is ...TI...ming, no, tim...ING, ummmm...timing. TIMING!
All we can do is put it out there in various forms and hope the wind comes along at the right moment.
Interesting perspective! Along the same lines, Claude would almost certainly fail in a PhD program. So I completely agree: Universities need to do more of the stuff that happens in PhD education and less general education.
The immediate thing that strikes me is Clause probably already knows everything that will be taught - it’s normal output is a post-graduate expert writing for a general audience. Its big challenge would be making sure it didn’t show too much advanced knowledge too early.
(Selective areas are fine - my CS project at 18 used a textbook borrowed from an older friend).
The other thing is the significant research in favour of students reading things on paper.
That isn’t an argument against AI teaching - it’s a good argument against both screen-based learning and in-person lectures.
I feel the danger is using AI to go further down an e-learning path that isn’t working - students reading AI summaries from Claude, all sides optimising for the minimum work to hit target.
What we need is AI as tutor rather than teacher - guiding reading, asking questions to check depth of understanding. This is clearly feasible, but not the helpful ‘personality’ the open market AIs have.
I teach one asynchronous online Lowe division gen ed course. While I agree on so many levels with your sentiment, there are always those students who enormously benefit: the busy single parent student, the student dealing with a chronic illness or mental health issues or being neurodivergent and even the extremely busy student athlete: they appreciate the flexibility and less social interaction so much, because they need it. I have definitely toyed with going back to in-person, but this always ends up outweighing my decision.
I am in biology. I hear you about farming out gen ed to AI, but I disagree that at least in biology a totally standardized curriculum would even be valuable. Biology touches on so many issues in medicine, environment and society; I try to be very thoughtful in making these connections and often students do work through complex issues like climate change, pollution, abortion, evolution etc. My work here is important and I take it seriously even when teaching majors might be more fun or rewarding.
Absolutely. But if a significant portion of the class are using AI to do the work, the value of the class declines for the people who need it. There is no current validation system to confirm a living breathing human did the work. That's my concern. I think we're in agreement about the big picture.
Yes, AND let’s say that K-12 education, and gen ed as a whole, would also be better if it centered human relationships and human development. (Seems like you agree? And maybe this post is mostly meant to reveal this obvious truth?) Claude also wouldn’t do well in any well designed K-12 environment or in any true gen-ed course.
It seems likely to me that the kind of undergraduate college experience you speak of will become the only kind that people are willing to pay for, and it will be expensive. What you speak of is what I think is accomplished today at Hillsdale College and perhaps at other colleges that I know nothing about. The secret sauce is "experience," as contrasted with to "knowing."
For years I have taught my students that if you have learned something, then you "know" something you did not know before, or you can "do" something you could not do before, or you have "experienced" something you had not experienced before. These are not mutually exclusive categories, of course.
I am a certified online instructor, as I have been for about 12 year, with certification from the Sloan Consortium; it was a high quality program. I agree with much of what you wrote. Asynchronous online courses are satisfactory for achieving the "know" category. Depending on the skill of the online instructor, and the intent of the online students, both the "know" and the "do" categories can be accomplished satisfactorily, and sometimes even fabulously. But the "experienced" category frequently goes unmet, particularly if the experience requires the presence and engagement with other humans. Writing in a discussion forum or discussing topics with others in a Zoom session is not a good substitute for the experience of growing up with a smallish set of real humans interacting in real life in an academical village, as it were.
I read an essay a few weeks ago that argued that college attendance in the future will become like fitness facility membership. The only continuing members have ulterior motives. They are not generally interested in an imprimatur; they are motivated to become or remain healthy and strong, for whatever reason. That sort of college experience is expensive, perhaps spectacularly.
Today, if I want to learn something that involves mostly "knowing" I turn immediately to Ai. If I want to learn something that involves a heavy dimension of "doing," Ai can be of some use, but not much. If I want to learn something that involves a heavy dimension of "experiencing," Ai is next to useless.
I personally think that independent certification, e.g., CPA, CME, CFA, etc.) will be the way to go and should naturally emerge due to creative destruction of voluntary exchange. I think that authentic evaluation for asynch online courses right now is nearly impossible without oral examination, which does not scale. Bryan Caplan makes the case ably, in my opinion, that signaling is all that has been going on for decades. I believe that reliance on a college's reputation has already become critically important to employers of corporations and other organizations.
That last line is the real kicker! It applies to many aspects of Industrial Age Schooling -- if it's only the "product" that a course is after, then Claude could get the grade! It's time to envision new roles for the teacher.
Years ago I asked a young man what he was reading. We discussed his Ivy League education and he revealed he had not let the university prescribe his path, but had gone into the Dean’s office to let them know what he wanted to learn. This was a revelation: that one could choose an autodidactic education! The article is wonderful, Hollis, and the comments by The One Percent Rule and Bill Benzon are encouraging to contemplate. Who would skip a class that was designed as desired, not just required? Sclerotic schooling is “so yesterday” and your proactive thinking is exciting!
How little Claude contributes to its fellow students! No friendship, no study sessions (testing one another), no devil's advocate: "can someone explain why this isn't just useless?"... Maybe Claude only offers rivalry?
I would hope that students enjoy a *good* course/ seminar series because each of them is part of the learning (I suppose it's still Socratic if the fellow student does this, rather than the "teacher"?)
Speaking of rivalry, does anyone remember how Anne (of Green Gables) studied her heart out to beat Gilbert? I can imagine someone hating/ resenting "Claude" for taking a human's place. Claude wouldn't benefit, though (as Gilbert did), so it would seem that human rivalry benefits more individuals...?
Color me sympathetic, Hollis. I've always thought that the push for "everyone gets a college degree" was a mistake. In practice it seems to me the result of this expansion is a lot of gen ed, poorly delivered, along with vocational training, also poorly delivered. So I agree, shift the gen ed to the high school level (which is the level it's at anyhow). Education beyond that should be direct and personalized, whether it's hands-on skills with a large physical component, or more abstract intellectual skills.
That's one thing. And then there's writing. Since my degree is in English and my one faculty post was in an omnibus department in a tech school (Language, Literature, and Communications at RPI) I've had to teach composition, one of the most thankless educational jobs in the world. It's always seemed me that writing needs to be taught the way musical performance is taught, one-on-one. That's how I learned to play the trumpet, and a bit of piano, between the ages of 11 and 17 (plus a semester of trumpet at Peabody in my early 20s). But assigning every student their own writing tutor is just too expensive. But it would get the job done, and all students need it, the good writers as well as the deficient.
Finally, a thought that's sort of tangential to your essay here, but I've been thinking about it. I've been hearing that students are using AI on their term papers etc. and "they're not learning to think!" (Duh!), as though learning to think was important to what colleges and universities actually do. In a system that's organized to assign and receive grades, and the grades are inflated (my graduating 3.3 at JHU back in 69 was well above average; now it's what, a badge of shame?) letting the AI get the grades is the thing to do, lowest possible opportunity cost. But it's clear that that smarter and more knowledgeable you are, they more effective you can be in using even the best AI. It seems to me that in a genuinely competitive environment that should matter. And if somehow you manage to get students interested in something, they learn to think even while making extensive use of AI.
People don't like being turned into IBM punchcards through industrial education, but if that's the system you put them into, they'l be punched and try to create some room for themselves elsewhere. Change the system, drastically.
Nailed it… I will never forget the interrogation I got from my Cornell faculty advisor when I converted my senior year courses from Chemistry to Electrical Engineering…
The value proposition that universities provide is (a) in person contact and (b) attendance confirmation. Claude would fail my finance classes leading to a CPA because I only do them in person and I take attendance (at the undergraduate level).
Kids are glued; AI is only another screen, though worthwhile.
On the K-12 level, high achievers who might most benefit from an AI tool need socialization skills. Schools often provide food, find abuse, provide community sports opportunities, and connect students with future employers. Claude isn’t doing that.
Hollis, this is a fantastic and provocative piece. The "Claude Test" is a brilliant diagnostic tool that perfectly exposes the vulnerabilities in the current higher education model. I agree, universities market themselves as bespoke centers for human transformation while often operating on an industrial model of content delivery, especially in the online, asynchronous courses that Claude could so easily ace.
If an entity with no lived experience, no capacity for growth, and no genuine understanding can pass a course, then that course is not assessing human learning. It's assessing information synthesis and pattern replication. Yes to this: a call for universities to double down on mentorship, community, and ethical development, the real "magic dust", is exactly the conversation we need to be having.
Where I would offer a counter-argument, however, is with the prescription to "jettison Gen Ed" and outsource it to AI.
I guess, the problem isn't higher ed as a concept, but its current, often impersonal delivery. By outsourcing it, we risk turning the foundational stage of higher education into a sterile, check-box exercise, as you suggest, thereby depriving students of the very chance for discovery that might shape their entire future. We would be conceding that the start of the journey doesn't need the "magic dust," when in fact, that's where it's often needed most.
Perhaps a better solution isn't to unbundle higher ed, but to re-infuse it with the human-centric values you champion. Let's use AI as a powerful tool, a "super-TA", to handle the rote aspects of these courses, freeing up faculty to run smaller, Socratic discussion groups, even for introductory material. The challenge issued by the Claude Test shouldn't be to cede this territory, but to transform it into something Claude cannot pass: a genuine exercise in human curiosity, discovery, and connection.
Is the complexity or advancement of knowledge *within* fields necessarily isometric with their teaching, though? An advanced seminar in history, about a very specific period or edge-case topic, might need less personal interaction than the 100-level courses where students are taught how to think like a historian, albeit with well-worn “content.” Unfortunately, departments often save their best teaching for advanced majors who have shown that they don’t need it, while the intro-level courses are considered low-status and menial at the supply side, and therefore regarded as such on the demand side. In other words, the personal transformation might happen at the start of learning a subject, rather than the end.
Also, to what extent is this also an assumption about *speed* specifically, and not just organizational efficiency? You can learn piano in a year, but 12 very good teachers can't get you there in a month.
Gen ed has been built with no theory of *mind* and no theory of *knowledge*—just liberal representational politics, but applied to academic disciplines instead of people.
I wish I could restack your note! It is the key point!
An interesting article ,as usual! It seems there are two connected issues here.
1.Why not do Gen Ed in high school,exploiting AI ? High school in the US is not a primarily academic enterprise ,with a few notable exceptions (ie Bronx HS of Science,etc) ,but rather a rite of passage.. Most intelligent,motivated students are bored stiff and basically marking time .There simply are not enough qualified teachers to meet the demand for AP and other such advanced courses.Doing Gen Ed by AI could rectify this situation.It could also shave a year or more off "college" and,more importantly, allow students to spend their college years exploiting the real resources of universities.
2.What to do in college? With mass produced Gen Ed out of the way, students could partake of advanced,specialized classes that actually utilize faculty expertise.More importantly,students could get involved in research and other experiences that feature personal mentorship by faculty ,thus leading to substantive intellectual and personal growth.
Yes this is the conversation we need to be having!
I took this month off from Substack (mostly) to read books, so just coming to this one.
It brought to mind something I found in my reading. Walter Lippmann explaining Henri Bergson's appeal: "He is popular because he says with splendid certainty what thousands of people have been feeling vaguely."
If an AI chatbot can earn a degree, then we need to make significant changes. I think thousands feel this vaguely. I hope you saying this with splendid certainty means you your way to becoming popular.
Thank you and welcome back! If you agree, I would love to make this point more and more forcefully and widely. The response has been a little tepid!
Well, we've been talking about that tough nut for a while, right? I think it all comes down to that old Steve Martin joke that the key to comedy is ...TI...ming, no, tim...ING, ummmm...timing. TIMING!
All we can do is put it out there in various forms and hope the wind comes along at the right moment.
Interesting perspective! Along the same lines, Claude would almost certainly fail in a PhD program. So I completely agree: Universities need to do more of the stuff that happens in PhD education and less general education.
https://clauswilke.substack.com/p/phd-level-intelligence-or-the-graduate
Love this and just subscribed to you -- thank you!
The immediate thing that strikes me is Clause probably already knows everything that will be taught - it’s normal output is a post-graduate expert writing for a general audience. Its big challenge would be making sure it didn’t show too much advanced knowledge too early.
(Selective areas are fine - my CS project at 18 used a textbook borrowed from an older friend).
The other thing is the significant research in favour of students reading things on paper.
That isn’t an argument against AI teaching - it’s a good argument against both screen-based learning and in-person lectures.
I feel the danger is using AI to go further down an e-learning path that isn’t working - students reading AI summaries from Claude, all sides optimising for the minimum work to hit target.
What we need is AI as tutor rather than teacher - guiding reading, asking questions to check depth of understanding. This is clearly feasible, but not the helpful ‘personality’ the open market AIs have.
The difference between earning a degree and getting an education has never been more obvious than it is right now.
I teach one asynchronous online Lowe division gen ed course. While I agree on so many levels with your sentiment, there are always those students who enormously benefit: the busy single parent student, the student dealing with a chronic illness or mental health issues or being neurodivergent and even the extremely busy student athlete: they appreciate the flexibility and less social interaction so much, because they need it. I have definitely toyed with going back to in-person, but this always ends up outweighing my decision.
I am in biology. I hear you about farming out gen ed to AI, but I disagree that at least in biology a totally standardized curriculum would even be valuable. Biology touches on so many issues in medicine, environment and society; I try to be very thoughtful in making these connections and often students do work through complex issues like climate change, pollution, abortion, evolution etc. My work here is important and I take it seriously even when teaching majors might be more fun or rewarding.
Absolutely. But if a significant portion of the class are using AI to do the work, the value of the class declines for the people who need it. There is no current validation system to confirm a living breathing human did the work. That's my concern. I think we're in agreement about the big picture.
Yes, AND let’s say that K-12 education, and gen ed as a whole, would also be better if it centered human relationships and human development. (Seems like you agree? And maybe this post is mostly meant to reveal this obvious truth?) Claude also wouldn’t do well in any well designed K-12 environment or in any true gen-ed course.
It seems likely to me that the kind of undergraduate college experience you speak of will become the only kind that people are willing to pay for, and it will be expensive. What you speak of is what I think is accomplished today at Hillsdale College and perhaps at other colleges that I know nothing about. The secret sauce is "experience," as contrasted with to "knowing."
For years I have taught my students that if you have learned something, then you "know" something you did not know before, or you can "do" something you could not do before, or you have "experienced" something you had not experienced before. These are not mutually exclusive categories, of course.
I am a certified online instructor, as I have been for about 12 year, with certification from the Sloan Consortium; it was a high quality program. I agree with much of what you wrote. Asynchronous online courses are satisfactory for achieving the "know" category. Depending on the skill of the online instructor, and the intent of the online students, both the "know" and the "do" categories can be accomplished satisfactorily, and sometimes even fabulously. But the "experienced" category frequently goes unmet, particularly if the experience requires the presence and engagement with other humans. Writing in a discussion forum or discussing topics with others in a Zoom session is not a good substitute for the experience of growing up with a smallish set of real humans interacting in real life in an academical village, as it were.
I read an essay a few weeks ago that argued that college attendance in the future will become like fitness facility membership. The only continuing members have ulterior motives. They are not generally interested in an imprimatur; they are motivated to become or remain healthy and strong, for whatever reason. That sort of college experience is expensive, perhaps spectacularly.
Today, if I want to learn something that involves mostly "knowing" I turn immediately to Ai. If I want to learn something that involves a heavy dimension of "doing," Ai can be of some use, but not much. If I want to learn something that involves a heavy dimension of "experiencing," Ai is next to useless.
All good points. I think the problem is validation. How can you tell that knowledge has been transferred. Was it possible to tell before AI!
I personally think that independent certification, e.g., CPA, CME, CFA, etc.) will be the way to go and should naturally emerge due to creative destruction of voluntary exchange. I think that authentic evaluation for asynch online courses right now is nearly impossible without oral examination, which does not scale. Bryan Caplan makes the case ably, in my opinion, that signaling is all that has been going on for decades. I believe that reliance on a college's reputation has already become critically important to employers of corporations and other organizations.
Agreed, which means what is left when signaling is over...?
Monastery? Institutes? The factory model of higher education appears to be a dead man walking.
That last line is the real kicker! It applies to many aspects of Industrial Age Schooling -- if it's only the "product" that a course is after, then Claude could get the grade! It's time to envision new roles for the teacher.
Years ago I asked a young man what he was reading. We discussed his Ivy League education and he revealed he had not let the university prescribe his path, but had gone into the Dean’s office to let them know what he wanted to learn. This was a revelation: that one could choose an autodidactic education! The article is wonderful, Hollis, and the comments by The One Percent Rule and Bill Benzon are encouraging to contemplate. Who would skip a class that was designed as desired, not just required? Sclerotic schooling is “so yesterday” and your proactive thinking is exciting!
How little Claude contributes to its fellow students! No friendship, no study sessions (testing one another), no devil's advocate: "can someone explain why this isn't just useless?"... Maybe Claude only offers rivalry?
I would hope that students enjoy a *good* course/ seminar series because each of them is part of the learning (I suppose it's still Socratic if the fellow student does this, rather than the "teacher"?)
Speaking of rivalry, does anyone remember how Anne (of Green Gables) studied her heart out to beat Gilbert? I can imagine someone hating/ resenting "Claude" for taking a human's place. Claude wouldn't benefit, though (as Gilbert did), so it would seem that human rivalry benefits more individuals...?
Color me sympathetic, Hollis. I've always thought that the push for "everyone gets a college degree" was a mistake. In practice it seems to me the result of this expansion is a lot of gen ed, poorly delivered, along with vocational training, also poorly delivered. So I agree, shift the gen ed to the high school level (which is the level it's at anyhow). Education beyond that should be direct and personalized, whether it's hands-on skills with a large physical component, or more abstract intellectual skills.
That's one thing. And then there's writing. Since my degree is in English and my one faculty post was in an omnibus department in a tech school (Language, Literature, and Communications at RPI) I've had to teach composition, one of the most thankless educational jobs in the world. It's always seemed me that writing needs to be taught the way musical performance is taught, one-on-one. That's how I learned to play the trumpet, and a bit of piano, between the ages of 11 and 17 (plus a semester of trumpet at Peabody in my early 20s). But assigning every student their own writing tutor is just too expensive. But it would get the job done, and all students need it, the good writers as well as the deficient.
Finally, a thought that's sort of tangential to your essay here, but I've been thinking about it. I've been hearing that students are using AI on their term papers etc. and "they're not learning to think!" (Duh!), as though learning to think was important to what colleges and universities actually do. In a system that's organized to assign and receive grades, and the grades are inflated (my graduating 3.3 at JHU back in 69 was well above average; now it's what, a badge of shame?) letting the AI get the grades is the thing to do, lowest possible opportunity cost. But it's clear that that smarter and more knowledgeable you are, they more effective you can be in using even the best AI. It seems to me that in a genuinely competitive environment that should matter. And if somehow you manage to get students interested in something, they learn to think even while making extensive use of AI.
People don't like being turned into IBM punchcards through industrial education, but if that's the system you put them into, they'l be punched and try to create some room for themselves elsewhere. Change the system, drastically.
Nailed it… I will never forget the interrogation I got from my Cornell faculty advisor when I converted my senior year courses from Chemistry to Electrical Engineering…