The Future of Entry-Level Employment in a Post-AI World

Welcome to another episode of AI Goes to College, the podcast that helps higher ed professionals navigate whatever the changes are going to be due to generative artificial intelligence.

I am joined once again by my friend, colleague and co host, Dr. Robert E. Crossler from Washington State University.

Rob, how are things in the great Pacific Northwest?

I couldn't ask for a sunnier, nicer day than today.

All right, well, let's get to it.

So today we've got three big topics we want to talk about.

The first one is what others are saying about AI in education, what kind of the general AI press is saying, want to talk about brain rot a little bit.

And an article that's gotten unfortunately, a lot of press that we ought to at least address.

And then our last topic is what people are seeing on the horizon about fall off in entry level hiring.

Rob, the first topic was inspired by an article from the Neuron, which is one of my favorite AI newsletters.

I'll put a link to how you can subscribe to it in the show notes and a link to the article we're going to discuss.

So, Rob, did you get a chance to look at the article?

I did, Craig.

I thought it was a great summary of really what's going on in higher education in the world of AI, how different people are reacting and responding to it, and with some good guidance on what are some of the best ways to approach a changing landscape.

Much like the Internet, there's a paradigm shift in the world of information.

This is even more so and happening much faster.

So I thought I did a pretty good job of capturing a lot of what I've been thinking and seeing in a lot of other articles I've been reading.

And you used one of my favorite academic words, paradigm.

Yeah, it's a really good article.

So the Neuron is a great newsletter and they've got a podcast that comes out sporadically as well.

But they are one of my go to sources for staying up with what's going on with generative AI, but maybe more importantly, kind of how it's being thought about and applied in different industries.

So they're not a higher ed newsletter, but that article covered a lot of ground.

It's really long.

I enjoyed reading it, but I did do it in two sittings because it kept going and it was very interesting.

Yeah, let's highlight a few things and again I'll have a link to the article in the show notes.

One of the things I really liked is they put a too long didn't read bit at the beginning, which kind of covers most of the ground.

One of the things is that if I had to summarize the whole article, at least a big chunk of the article, it's all kinds of students are using AI now and they're mostly using it wrong.

I would add to that many faculty and educators are not being properly trained in how to use this in ways that are meaningful and purposeful in education as well.

So the two things make a very interesting combination of issues we need to get over.

That's where we come in.

We can be part of that training.

Most aren't being trained at all.

It's pretty rare to have any kind of real training available.

And I will put in a plug.

Either of us or the two of us together are certainly willing to talk with anybody who wants someone to come in and demystify generative AI for their faculty.

We're both really good at that.

We've been teaching non technical people how to leverage technology for a very long time.

I've been doing it for 40 years and I know you've been doing it a long time, Rob.

So if you need that, Craig at AI goes to college and we will get you hooked up.

I'm a little bit surprised and a little bit not surprised that we don't seem that much further along than we were a couple of years ago.

I mean, is that the impression you're getting as well?

Yeah, there's pockets of people who have gone all in and have learned how to do this and to do it well that are experimenting and are playing.

And this article talks about those sorts of people as well.

But they are the minority.

I think we have a number of people who are trying a little bit of things that are very comfortable in the way they've always done things.

And just doing a little bit is a big stretch for them.

And then you have a fair number of people who ban AI flat out want to figure out ways going back to bluebooks, a drastic increase in bluebook sales.

If you don't know what blue books are, they're books with the blue cover on them that have 8 to 16 pages in them.

And students would show up and with a pen, you'd hand them a blue book and they would write out answers by hand.

And this is an approach that saw a drastic increase of how people are addressing AI is by saying, hey, let's remove technology from the equation and see what students know.

Yeah, I read somewhere where it's like an 80% increase in blue book sales.

I think our friend and colleague Tom Stafford has accounted for a Chunk of that.

That's not an unreasonable approach.

One of the things that we've talked about this before, separating out learning from assessment test is about assessment.

What have you learned?

The learning activities that we often grade really are only kind of about assessment and may not be necessary.

So I want to be clear, at least I'm not taking a really dim view of going old school to kind of make sure that students are learning what they need to learn.

Do you agree or are you on a different page there?

A little bit.

I think depending on what you're doing in the learning process, that might be an appropriate way to do assessments.

But I've been reading more and more that the way we assessed historically was already getting outdated.

It was already getting fairly easy to game for students who could game things.

And it moved to, you know, it's really been moving towards, let's do different assessments that don't really focus on the object you created, but more the process of creation.

And I think if we start wrapping our heads around what does it look like to assess your ability to engage in critical thinking in a learning process, regardless of the topic area, it changes how we would assess that because it's a lot harder to do that with our traditional methods.

Let's get into that a little bit.

One of the things about this idea of brain rot, which we're going to get into a little bit more in a minute, is that students don't do the hard work of learning.

And I think if we can find ways for students to demonstrate that they're doing the work of learning, that might be the way to go.

I think that's kind of hard to do, especially at scale.

You know, a lot of these problems are much easier in a classroom of 20 or 30 than they are in a classroom of, you know, 90 or 100 or more.

So I think we face some difficult choices there.

But I think you're right that if we can assess the process, but let me push back a little bit on that isn't what they know or what they know how to do.

The important thing I know the journey is the reward in life, but in school, it's like, you either know this stuff or you don't.

True.

And I think when in a world where I can do a Google search to get information and I can then give it to you.

Okay, now I demonstrated that I. I know something because I created something through the help of Google, even if I borrowed it from Wikipedia.

And then I moved that into the AI world where I just give it a prompt and Create something and give it to you.

Demonstrating that I know something, I don't actually know anything.

And there's more and more studies that are showing that when you do that, you actually can't recall any of those sorts of things.

But yet there's huge value in being able to very efficiently create documents and to do things.

So more important than the creation of that artifact is how to do so in a critical way, where you're engaged with the information and bringing a human element that is critical of what's going on, that is shaping and forming it, to use it as a tool to help you, as opposed to a tool to do all the thinking for you.

Well, I think you've hit on a really fundamental point and one that we're going to have to work a lot on collectively, and that's rethinking what we want students to know.

So, you know, back in the day, you memorized formulas and facts because if you didn't know it, you had to go through a lot of work to go find it.

I'm old enough to have gone.

Had to actually go to the library.

I mean, really go to the library and look something up on, you know, the Dewey Decimal catalog cards and that sort of thing.

Well, you don't have to do that anymore.

And with the advent of the web and Google, you know, memorizing different facts.

Do you know when the Battle of Hastings was?

No good.

Could we find out when the Battle of Hastings was?

Ask Siri.

I'll know it.

Yeah.

And, you know, if you're a historian, you ought to know kind of the general range, because that gives you where it fits in the flow of historical events.

But, you know, the exact year, the exact date, look it up.

I agree.

And one of the things, I think that is the scary thing about generative AI and where we have to help our students know how to use it, we have to bring that critical thinking in, is I was reading an article the other day that in Russia, they have 3 million articles being created every day about what their view is on the war with Ukraine.

And they're pumping it out in such a way so the generative AI technologies pick it up and believe that the Russian view, propaganda of the war is factual.

And then when that gets pushed out to AI solutions, if we trust it and say, oh, AI says this.

This must be the way things are are.

We have missed that ability to say whether it's, you know, a Google search that, oh, this is someone's blog post versus this was written in the New York Times.

Or somewhere where you might give a little more credibility to the source of that information.

And so the world of Google, in fact, you know, has its own issues, but you can vet it in ways that help you to decide if it's reliable.

In AI, from a propaganda perspective, it can be easily manipulated.

No, that's right.

And in our textbook, you know, we have a big chunk on information evaluation that gets into exactly how to avoid falling into that, that low quality information trap.

But I think the world changes and we've got to change along with it.

And yeah, there are certain facts that and certain skills that you just need to be able to do in the moment, depending upon your field.

You know, if you're a nurse in an emergency department, you know, you don't have time to look up every little thing you've got to do for every patient and every profession has its parallels, but for a lot of things, if it takes you 30 seconds to look it up, it takes you 30 seconds to look it up.

Who cares?

I wonder about spreadsheet skills, which have been our go to for a long time.

I mean, right now generative AI is kind of okay on a lot of the spreadsheet stuff, but it's not long before you'll be able to say, here's a spreadsheet, do this and it'll just get done.

I know Copilot can kind of do some of that, although I never have.

Good luck with Copilot.

Well, one of the things I have enjoyed with Copilot and Excel is when I do a function, I can't quite remember it.

I used to go to Google and within five to 10 minutes I could find a place that helped me to do it.

Most of the time I can get pretty good guidance from Copilot that allows me to implement exactly what it does.

And it's removed Google from that equation, which has sped up the process by a few minutes.

Yeah, Gemini is great for that.

I always forget the function to separate out given names and surnames to left something, something.

But with Gemini or Copilot you can get it in 10 seconds.

Yeah, but I think we need to really reconsider what we want students to learn in a world that's going to be driven by AI.

And I think if you're preparing students that are going to go into any kind of knowledge work, AI is going to be a part of it.

Matter of fact, just before the recording session was on a call with a guy that runs a group of CFOs that I talked to a year ago about AI.

They want me to come back again to give an update and to talk about practical applications of AI.

And so it's just going to be part of the world of knowledge work.

And we really have to reconsider what we want our students to learn and to be able to do so.

Craig, one of the things in this article that gets to this a little bit is a different way of thinking that's about mastery as opposed to the amount of time students spend in a space.

So in higher education, a lot of universities require for a three credit hour class, you have a set number of contact hours, usually about three hours during the week for that three credit hour class, and then an expectation that maybe for every hour in class there's two hours outside of class that students are working.

It's very much a time based formula.

And prerequisites for classes or to get credit for graduation, you usually need a C or better, which means across the board we're okay with a student learning 75% of the material deemed necessary in the assessments.

The alternative is a mastery based approach where we don't have students move on to the next topic area until they demonstrate they have mastery in the topic area that we had determined was necessary before they moved on.

And so if you can do that faster, you may actually advance quicker and be done in a shorter amount of time and not have to engage as much as if you're someone who struggles with those concepts.

Maybe math is not your strong suit, maybe you spend longer to develop that mastery.

But then when you move forward, you can rely on the fact that you know that.

And in this world of generative AI, where so much is changing, one of the arguments is, is maybe that mastery approach is going to help us to individually work with students using the generative AI tools or whatever that allows us customize things in a way that pushes things forward.

Is that possible, Craig, in your experience in higher education, to move to that sort of paradigm?

Because they seem to be such opposite mindsets of how to do education.

Is it possible?

Yeah.

Is it going to be easy?

No.

I mean, I think individual programs can move in that direction.

I mean, it makes total sense for us in information systems to take that approach because employers don't care about random facts.

They care, can you do this or can you do that, which is mastery?

I think it makes a lot of sense for those kinds of programs.

Does it make sense in art or music?

Maybe, maybe not.

I don't think there's going to be a one size fits all, but something you touched on is going to be absolutely key.

And that's leveraging AI to make this manageable.

And I'll use the term again at scale, because it is possible to have those individualized learning pathways.

I mean, it's not easy to do, but it's certainly possible.

And that really lets the students who are more talented in that particular area go further without leaving behind the ones who are less talented.

You and I both teach core classes from time to time where you've got some really good students that are really into the topic and they could go to the moon.

And then you've got students who either aren't very talented or maybe aren't as into the topic and you've got to teach somewhere to that middle.

And everybody has to do this.

In any kind of a core class especially, that could go away.

But I think the way to do it is in individual programs.

Rob, you're a department chair.

You could start pulling this off in individual classes where it makes the most sense.

I'm not sure it makes much sense in the junior level.

Everybody in the business school has to take a class, but your coding classes, your database class, you know, those kinds of more skill based classes, totally feasible.

So here's the question for you as someone who's been in that dean's role and has seen the higher level.

When the university has policies that dictate the number of contact hours and how that makes up a class, how do you implement a mastery based approach when the system's rules aren't designed to allow for that?

Do you ask her forgiveness rather than permission?

Yeah, what, what you could do is push that long tail pretty far out.

Where it's not you get to this particular level of mastery and you're done.

It's you keep mastering more and more and you're done at the end of the term.

So I think that's a little bit more work, maybe a lot more work.

But with a coding class, with a computer programming class, you just have harder and harder problems and they just keep going until the term is over.

Although I think a lot of that, you know, that 40, whatever hour, contact hour, I don't even know what that means in an online class.

So I think a lot of that kind of stuff may technically be required.

And if it ever got audited, you'd want to have some story to tell.

But it's not like it was back in the day.

I think what I would do to kind of get this transformation going is pick my battles and start chipping away.

So if you do it in One class.

Now you've got a proof of concept, makes it easier to sell.

Faculty and administration on the second class and the third class and the fourth class and doing a whole program that way, pretty heavy lift taking.

I mean, what do you have, six or eight required classes.

Three MIS classes that are required and then a couple of electives.

So doing that for three or four classes in a program, I think over the course of a year or two, that's totally feasible.

So anything else on this article?

But we're going to come back to it because it applies to some of the other things we're going to talk about as well.

But I think we, we hit on most of it.

I think that whole concept of mastery is one that is worth wrestling with.

Yep, yep.

And again, to all of you listeners out there, I really would encourage you to, to read this article and to subscribe to the Neuron.

It it's a solid newsletter, although now I won't be able to look as smart when I write stuff that nobody knows about if everybody's reading the Neuron.

All right, let's move on to the next idea, the next topic which is brain rot and cognitive debt.

This is an article by Nick Potkalinsky.

I'll put in a link in the show notes about this as well.

But I thought this was a pretty good article that plays off of an MIT working paper that's received a lot of press unfortunately.

So Rob, you want to start on this one?

Yeah.

This is interesting.

First, I'm going to caveat this with this one study had a very small sample size and so I want to be careful that we don't take too much away from this because it needs more scientific rigor to establish what they're saying.

But I think what they're saying has merit and what they did is they actually connected things to the brain to see what sort of activity was going on.

When people were writing essays using various different tools and people who use Chat GPT and just straight up used it had very little brain activity going on compared to those people who were writing things manually or doing things with other technology tools like search engines helping them.

And really what they show is then recall of what people created wasn't present for people who had just used Chat GPT but people who took the other approaches could remember what it was that they had written.

So not terribly surprising that if I copy and pasted something and didn't really think about it that I couldn't remember what that thing was.

Yeah, I'm having to filter out what I really want to say about this because this is a G rated podcast, but okay, if I go to the gym and get a machine to lift the weights, I'm not going to put on muscle.

You know, the methodology is really interesting and much of what is claimed in the media about this article is not necessarily what the authors claimed.

This article also has not been peer reviewed, although the last author, Patty Mays, who I think runs this lab, she is a legit human computer interaction guru.

I was citing her work as a doctoral student.

So, you know, this is a very solid study, but it hasn't been peer reviewed.

I think over all of the experiments they did, they had like 50 something people.

The ones in some of the conditions were, you know, the sample size went down to like eight or nine.

So it's hard to draw a lot of conclusions about this, but it's kind of, yeah.

So wait, if I write something and put some work into it, I can tell you what I wrote about a week later.

If I copy and paste something.

I can't, right?

This is nothing new.

If you think about people's ability to remember phone numbers, I have about three or four phone numbers that I actually know the rest.

I rely on my phone to remember that information for me.

And I don't think that's necessarily a bad thing.

Unless my phone gets stolen all of a sudden I can't, you know, get a hold of someone to help me.

But when you get arrested, how are you going to know who to call to throw your bail?

I know, right?

There you go.

So.

No, that's true.

In fact, I just wrote down important numbers on a little three by five card for that very purpose.

You know, if your phone dies or, you know, I mean, I couldn't tell you my brother's phone number off the top of my head.

And I talked to him a couple times a week.

You know, it's the way it is.

But like you said, I'm not so sure that's a bad because we've got limited cognitive capacity.

And if that's something I don't have to chew up part of my brain to remember, you know, frees it up for something else.

Of course, it could be nonsense that I put in there instead.

But what I think is important and where I hope some more studies go is let's establish the extreme, the copy and pasting from a generative AI tool.

I remember way less than if I did all the research and put the thoughts together and wrote it myself.

Those are the two extremes of this experimental condition.

What I would love to see in more experiments is what level of engagement with the generative AI results result in the best learning and actual knowledge retention.

For me, that's valuable.

And what does that begin to look like?

And I don't think we know perfectly what that answer is, but I think it's more than copy and pasting the results.

But what are some of the proper ways to engage with what you receive in these queries?

So that way, A, you retain what you learned and B, you have a level of confidence that what is created aligns with your thinking, is consistent with the facts as they are.

And so to me, that is where I think some really interesting work could be done.

Yeah, absolutely.

And Nick and the authors of the MIT study used the term cognitive debt, which somehow got translated into brain rot in the press.

And basically what that means is that if you just copy and paste from ChatGPT or whatever, it's kind of a mental IOU.

And so, yeah, that helps you get this part done more quickly.

But if you have to recall that information later, you've got to repay that cognitive debt because you've got to go back and look it up or, you know, however you want to do it.

There's also very little about the long term effects.

This is a big problem in higher ed and one that we really need to get a handle on.

If students are copying and pasting or some equivalent of that throughout many different courses in their curriculum, they aren't going to learn these things.

They're going to have a really heavy cognitive debt to pay.

You know, one class, one assignment, it's kind of like eating one cookie doesn't do anything.

Eat a box of cookies every night, not so good.

So, Craig, one of the thoughts I had around this is if you think about the value of higher education, right, a degree from Louisiana Tech says something about what the student has accomplished.

And it's a symbolic of the learning that occurred.

Washington State University degree has meaning because we ensure that students have learned certain things.

It's why we have academic integrity of a student's caught plagiarizing if it happens enough times, are expelled from the university, all of these things are done.

So that way when someone sees, oh, you have a degree from that university, there's meaning and there's value there.

And that's with this whole idea of cognitive debt.

If we allow that to accumulate through a degree program and you successfully graduate, what do the employers then view as the value of the degree from that institution?

And so I think this is really important for us to get a handle on.

Because if we get a handle on it and we can ensure that students are not graduating with that cognitive debt, that means the value of the degree they got from our institutions still is worth the price paid.

Right?

Well, and I want to make two follow on comments.

One is the schools that do this really well figuring out how to leverage AI, their value is going to go up.

It just is.

The other thing is that one of the kind of overarching messages of all of this is learning is work.

If you don't put in the work, you don't learn.

But there's a parallel to that.

Educating students is hard work.

And I think in a lot of ways we've gotten lazy, you know, multiple choice, machine graded tests.

And I'm guilty of that.

It's hard work and we're getting pushed into needing to do some of that hard work if we want to have continued long term success for our institutions or our programs.

So go ahead.

Sorry.

Yeah, I was going to say one of the things, you know, as individual faculty members, we can see this and we can say, yes, I need to be doing that to ensure the credibility of the course that I offer.

But institutionally we need collectively to get that sort of shift happening.

How do we encourage air quotes around that faculty throughout a department, throughout a college, even throughout the entire higher ed institution to put in that hard work to begin shifting to ensure that this cognitive debt doesn't occur?

That's a tough question.

So let me give you a couple different reactions to that.

One is that if you can convince them that doing so helps them achieve some higher purpose they've got, it'll fall into line.

Long time ago, I wrote a little article for the Flagstaff Business News called something like the Secret to Herding Cats.

And the gist of the article was this.

You know, you've heard the herding cats metaphor.

It's pretty easy to herd cats.

Just put some food out.

If they're hungry, they're going to go to the food.

And so you can do the same thing with people.

If you can show them how doing what you need them to do aligns with something they care about, then they're much more likely to do it.

And of course, you know, this isn't going to work at the margins.

You know, we're assuming that faculty care about their students and most in my experience do.

But as a practical matter, I think it's gotta be a small wins approach.

Pick a couple of classes, pick a couple of faculty members, get them doing it.

They can help others do it, others will see and it'll kind of spread.

I mean this kind of started.

It kind of went the same way with websites for classes.

You know, a few faculty started creating websites and now they don't have to hand out study guides or they don't have to do this or they don't have to do that.

And then somebody says, hey, can you show me how to do that?

And it spreads and spreads and spreads until it becomes institutionalized.

And, and I think we're going to see the same sort of thing with AI.

The difference is the stakes are a lot higher this time.

The second thing I wanted to bring up is in Nick's article, once again link in the show notes, he makes a number of suggestions on what we ought to do to kind of avoid this cognitive debt.

And you can read the article for all of them.

But the one that I wanted to highlight is visible metacognition because I think you could apply this to a lot of different kinds of learning activities.

And I'm just going to read from the article.

Ask students to annotate how and why they accepted or rejected ChatGPT suggestions.

Reflection Re engages executive networks.

So if you think about something and reflect on it, you're going to learn it better.

And so if you say okay, use ChatGPT, but you've got to go through and, and talk about why you took this answer from CHAT GPT and not this other answer.

Now lazy students can do this in a half assed way, but I think that's a pretty good approach that could be used maybe not for every learning activity, but for a lot of them.

Yeah.

And I'm taking that approach this semester, Craig, where especially the class preparation activities.

I don't so much as want the solution.

I want to see the process they went through to get to that solution.

And if you used AI tools, great.

Tell me how you did it.

Tell me how you critically evaluated it.

And that's going to be a required piece of what's submitted and it'll be part of the grading criteria for how they come to class prepared to engage in a conversation.

Yep, that sounds like a good idea to me.

All right, now I want to close this segment going back to the Neuron article they bring up Simon Sinek.

Have you heard Start with why that had a real impact on how I tried to lead as a dean.

By the way he talks about the excruciating journey.

That's the cognitive work that builds and strengthens neural pathways.

And in cognitive science they call it desirable Difficulties.

And so this really goes back to how I especially help educate doctoral students.

And on the Excel portion of my undergrad class, I tell them, I want you to struggle enough, but not too much.

You know, if you're not struggling at all, there's a problem.

If you find yourself banging your head against the wall for hours, that's a problem.

So we want to shoot for that golden mean that has these desirable intentional difficulties built into what we do.

So, again, both of these articles are really fantastic.

Have I mentioned that there are links in the show notes?

I don't know if I brought that up.

I think you have once or twice.

And one of the things that I think we haven't talked about, Craig, that's important as well, and it deals with issues from the Neuron article as well as the brain rot article, is what does it look like when we begin taking a Socratic approach to the classroom, where we're asking those why questions and really engaging with the fundamentals of that learning through dialogue and conversations, as opposed to death by PowerPoint as an extreme difference?

We've been doing Socratic dialogue since Socrates.

Again, BC something 100 BC.

We could look it up if you want to know.

Yeah, because it works.

The problem is scaling.

But again, this is where AI can actually be a huge benefit.

In fact, I built a little custom GPT type thing that did Socratic dialogue about moral decision making, and it wasn't that hard to do and it was pretty effective.

So that's the kind of thing we could do, is put this up at scale and you can get meta with this.

I mean, actually, it wouldn't be that hard to have AI do an initial evaluation of the students reflections on how they used AI.

So, all right, I feel good.

We've got paradigm in, we've got meta in.

What's our.

This could be a drinking game, you know, Bingo.

With Craig and Rob.

That's right.

Okay.

All right, so let's get to our last topic, and that is an article that I wrote that just came out today in the AI Goes to College newsletter about the rising entry level and unemployment and the coming skills gap.

So let me give the gist of this and then you can give your reactions.

Rob.

So basically, there's growing evidence that unemployment is rising among entry level workers, even for entry level workers who have a college degree.

And there are statistics in the article that you can look at if you want to, but it's noteworthy and it's increasing the gap between the overall unemployment rate and the unemployment rate for new college Graduates is widening.

It's hitting the tech sector pretty hard right now, really.

Any kind of entry level knowledge work is in jeopardy.

It's hitting primarily the tech workers right now.

But that is going to change.

That's because AI coding tools came out really quickly and were pretty good, according to a lot of people.

But it's coming.

But even in health care, it's coming.

So that presents a real problem for us as educators, you know, because, look, I don't know about you all, but if I'm trying to convince somebody to come to Louisiana Tech, I talk about our 90, like 96% or 94%.

I don't know what the exact number is, but it's, it's in the mid 90% placement rate for the college of Business.

You know, to a parent, that's a pretty good statistic to throw out.

What if that's 58% or 72%, you know, all of a sudden?

Yeah, not such a big selling point.

But it's also a matter of doing what's right for our students.

You know, not everybody, but most students go to college to launch successful, meaningful careers.

If the entry level jobs are cut off, where does that leap?

So, Rob, what do you think?

Yeah, this is something that I think about a lot, especially having a child entering their last year of higher education pursuing a data science degree and wanting to see them launch.

Right.

So it becomes very personal when I look at my own kids.

And part of me thinks this is an AI issue, but I've been reading some other things that suggest that there are alternative explanations as well.

So I hope that maybe there's a silver lining we can get to.

But where, where I've come down on this, as I think about this, is this reinforces the point that we in knowledge work disciplines need to ensure that we're teaching our students how to accomplish those skills with AI as a tool.

And the students who can articulate how AI has helped them to accomplish, whether it's finance skills or coding skills, are going to differentiate themselves from students who can't articulate that.

Preach it, brother.

I want you to say that again because that's really the key to all of this, I think.

Yeah.

So we need to establish and help our students to be able to articulate how they can utilize AI skills to accomplish the knowledge work so they're able to differentiate themselves from those students who can't articulate that and show that they have those skills.

AI is a tool.

We need to make sure students are prepared to use it in a workplace that is still figuring out how to use it and how to do some incredibly awesome things with it.

Absolutely.

Yeah.

And that.

So this is a macro micro problem.

I think higher ed's got some real challenges ahead at a macro level.

I care about students generally, but I really care about my students and students in our programs.

I can't do much about the macro level other than this podcast in the newsletter, but I can do something about the micro.

And I think, you know, we've been talking a lot about what can individual faculty do.

Individual faculty have that under their control.

They can start learning how to use the tools themselves and helping students learn how to use the tools, not only to produce the kind of work that they're going to produce when they're out there in their profession, but also to be able to learn effectively with the tool.

Because if you're not using AI to help you learn tools or help you learn things, you're really missing out.

I've got an ungodly number of deep research reports, many of which have nothing to do with our field.

You know, they're just about this or that, just something I'm interested in.

So I, I think you're spot on.

So let me get some alternative explanations.

Craig.

So it's not AI, Is this evil thing that's potentially causing our students not to get jobs?

Another thing that's going on in the economy right now, that's an important part of this as well, is there's a lot of uncertainty in businesses around the effect of tariffs and some of the other policy changes that are going on.

And what it's causing is a fear amongst existing workers to change jobs because oftentimes that's what opens up jobs for entry level positions is people advance their careers by moving.

And we're not seeing a lot of movement between jobs right now because people just are unsure and uncertain because there is just a lot of stuff going on beyond generative AI.

And so we may actually see things get shaken up as the older workers.

Right.

Those have been out 5, 10, 20 years, feeling comfortable, moving roles again.

Organizations that perhaps have put a freeze on hiring because they want to figure out the new lay of the land.

Once they figured it out, they may open up those positions and start hiring again as well.

And maybe in a way that's beginning to take advantage of AI and to put those tools in place to increase their opportunity to gain market share and to do different things.

So I think there's just so much going on.

I want to make sure we're not saying AI is The big boogeyman in the room.

And the world is a multi causal place.

But although that makes sense, the gap started before the tariffs.

I'm sure that's got something to do with the current statistics.

And the tech sector had some kind of employment rationalization post Covid.

You know, they ramped up a lot when we all went on zoom and all that kind of thing.

And so they needed to kind of get their workforces back to the right level.

But AI is a big part of it.

That's not going to change.

So, yeah, I mean, I think there's a lot going on just because the economy is a really complicated thing.

But.

We still, we can't.

And I know you're not doing this, I want to be clear, but we can't just say, oh, it's these other.

It's the tariffs.

I've read that.

Oh, it's the tariffs.

You know, tariffs are part of it, but they're not all of it.

Well, and I think those companies that figure out how to leverage AI by hiring workers that know how to use AI are going to be those companies that see increases in market share, that see successes going forward.

So given the landscape of where we're at, I think everything points towards taking efforts to prepare yourself for a world of AI.

And what I'd add to that too, you know, we've talked a lot about how do we get faculty to empower and enable students to do this.

I would speak to any student out there who might be listening to this podcast and encourage you to play with AI, to come up with creative ways of doing various different things and be able to tell that story.

Even if your faculty members aren't paving that path for you and showing you what to do, there's nothing that stops you from becoming that expert in the room through self learning.

Well, and for anybody who's got kids in the car right now, fast forward about 30 seconds or a minute.

You know, I read comments from students online that are saying, you know, our schools aren't teaching us anything about AI and so we're not going to be prepared.

That's a excuse.

You know, we should be preparing our students.

But look, you're grown adults.

Take some responsibility.

If the school's not going to do it, do it yourself.

If your school's also not teaching you how to cheat at Fortnite or, you know, whatever the game, I don't know, I have no idea what the games are.

Now.

That's one of two I can recall.

But a lot of kids are figuring that out on their own.

So you can do this.

Your school should be doing something and if they're not, they're not doing their jobs.

But you got to do your job too, as a student and that's not to rely just on your professors and your school.

Let me tell you a little trick.

If you go to Chat GPT and you ask chat GPT, how do I teach myself how to use generative AI in insert discipline here, it will probably give you a really good starting point and they can be a tool to help you to learn how to use the tool.

Oh, Rob goes meta a second time.

All right, so the last bit of this last topic, that's something I'm calling this Coming Skills Gap.

And I've been thinking a lot about this.

I mentioned it at a conference a few months ago and a lot of people think this is going to be a problem.

So we have entry level employment, right?

And as you said earlier, Rob, people get that four, five, six years in and they move to higher level positions, right?

They become managers or they go to a higher skilled position or something.

They might do that within their current company, they might leave, but they move up, right?

I mean, that's what we've always done.

What happens if you don't have the entry level people?

Who are you going to move into those higher level positions that, by the way, are going to be harder to replace or even augment in a significant way in the medium term with AI.

AI is not going to be great at strategic thinking.

AI is not going to be great at managing and motivating people, not anytime soon.

So I think we're going to have this chasm where there's just going to be nobody to fill in these positions.

And we've seen a little bit of this before, but I don't think we've seen it at the scale we're going to see it at.

So tell me I'm wrong.

No, I don't think you're wrong.

And I think this is a place where engaging with industry as people who recognize this is coming.

And I'm sure that there are certain people within industry that see this as a potential problem as well.

How do we work together to ensure that we pivot in the right way to where we are meeting those needs and we're going to try and fail.

We're going to see some experiments on how to set these things up.

But this is where us in higher education can't sit in the ivory tower and think that we know everything.

This could be again, a big word for Craig, a paradigm shift in that engagement of what industry expects from higher education and what higher education can be doing to help our students succeed as they enter into this changing market marketplace.

Yeah.

Yeah.

But any of you industry folks that are out there, you're going to have to do your part as well.

So it's going to be interesting.

All right.

We are pretty much out of time, Rob, any last thoughts?

Nope.

I think it was a great episode today, Craig.

I look forward to seeing it posted.

All right.

And so just as a reminder, you can find all things AI goes to college@aigostocollege.com you can email me at Craigi goes to college.

And Rob, Rob Crossler, C R O S S L e r@AI goes to college.com There's a contact form that you can access from aigostocollege.com Again, we are happy to help you in any way you can.

To quote a movie from long ago, it don't cost nothing just to talk to us.

So if you are interested, get in touch with us.

We're glad to.

Oh, God, here's another phrase I hate to use.

Jump on a Zoom call and talk to you about what you might need and how we might be able to help.

All right, Rob, anything else?

Nope.

All right.

Enjoy the cool weather.

I will be sweating and mowing this weekend.

All right, thank you.