Welcome to another episode of AI Goes to College, the podcast that helps higher ed professionals figure out just what in the world is going on with generative AI.As always, I'm joined by my friend, colleague, and co host, Dr. Robert E. Crossler from Washington State University.Rob, how are things out there?

Things are going well, Craig.We're past the halfway point of the semester and it just seems like the.

Process marches on and we're in strange days here.We've had continuing drama around a presidential change, but more importantly, we've had such a big drought that I'm going to have to mow my pond.Not the edge of my pond, the pond itself is so dry.But nobody cares about that.So let's get into the discussions.We've got three things we want to talk about today.And then at the end, we've got a couple of little AI tips that we might try to work in as well.All right, so, Rob, I sent you a couple of reports that just came out recently.One from the National Bureau of Economic Research, which is a well known, I don't think tank around economic issues, and then one from the Census Bureau of the United States, their Center for Economic Studies.And of course, there will be links to both of these in the show notes.They came to different conclusions.They were both looking at entry level hiring and whether or not it's been affected by AI.The NBER paper kind of said, no, not really.And the Census Bureau paper said, well, it depends.So, Rob, did you get a chance to take a look at these?I know I sent them to you at kind of the 11th hour.

Yeah, I did look at them and it was interesting, but not surprising.I think what we see in the data is it's really messy and it's hard to know exactly what's happening.But as I read between the lines a little bit, what I see is that the marketplace is changing, as it always changes, maybe a little bit more rapidly than it has in the past.And it begs the question for how are we paying attention to it and how are we ensuring that students are prepared for what is a changing world right now?

Right.Have you been reviewing my papers?Because I get interesting but not surprising in a lot of my reviews right before the paper gets rejected.

I think the AI models that are being trained to do reviews are using our podcast to train themselves.

Yeah, I think I would agree with everything that you said, but the details are kind of interesting.So I'm going to see if I can explain kind of how these work, but they're very Economics papers almost to the point of being incomprehensible.So basically the NBER report looked at unemployment averages across a number of years, I think 2024 to 2026.And they saw some seasonally adjusted differences, but pretty small.In the summer, the unemployment averages were 7.3% in 2026 and 7.8 in 2024.2025 Was kind of in the middle, but those are such small differences that they kind of don't mean anything.They're kind of high, higher than we'd like, but, you know, not big changes.And so their conclusion, that is basically that there's not really much evidence of an AI caused change in entry level employment.Does that kind of track with what you found?

I do think there's nuance in that that becomes interesting, but we can't have nuance.

Nuance, sorry.

I think it really does depend on.We go back to where human value is in what new hires are bringing to the table and how prepared they are for bringing, I've heard them called durable skills, soft skills and those sorts of things to the equation.That entry level jobs with people with those skills, AI is empowering them to do things they could never do before.And so that's where I think the messy is, is in some ways, jobs are opening for people that maybe they weren't available based on the skills they bring, where it may be at the expense of some people that had different skills.So it's kind of that ebbing and flowing of what the marketplace looks like.

Yeah, I mean, the differences could be washing each other out, but we really don't know that from this particular report.Unless there's something buried in there that I either missed or didn't understand.

No, I didn't miss it.And then that's where I think this is consistent with some other things that I'm seeing in my LinkedIn feeds and other feeds that I look at is that things look fairly normal, but it just very much depends on the part of the country or the world that you're in.What is it that they're looking for and what is it that you bring to the table?And so it's, it's a messy change that we're going through.So it's macro looks one way, micro looks another way at times, but.

And it's a big but.Don't say it, Rob.There's another piece to this picture, which is why we wanted to talk about these two reports together.So the CES report broke its analysis down by occupation, and what they found was Alarming.Particularly alarming for those of us who teach in the computing fields.So they broke down these different occupations in deciles according to how much unemployment entry level employment rather has fallen.And in the Most exposed the 10th decile, the initial employment falls an estimated 5 percentage points, which is bad.If you look at networking and telecommunications, that fell by 7.1%.It's a concerning picture and numbers are always great on a podcast, but I want to give you a few of these.So keeping in mind these are differences between whatever I'm getting ready to mention and the average, here's what the numbers are.Computer networking and telecommunications -7.1% computer science -6.1% computer and information systems general -5.4% computer programming -2.3.Management information systems and statistics, which is a weird combination.-2.1% Finance -2.1.Marketing -0.3 and accounting -0.1.So none of that is good.But that's alarming and a little bit surprising for the computer networking and telecommunications.But I think there's a hidden story here.I think we're seeing the canary in the coal mine.And let me tell you what I mean by that.Because of the nature of the training data in terms of professions, one of the first areas to be affected is computing, because the AI companies could go out and grab all of these existing code bases off of GitHub and other places and do a pretty good job of coding.It's a little bit harder to do that in something like marketing, but I think it's coming.I think we're seeing the leading edge of something that's going to affect all of knowledge work.We're seeing effects in journalism and some other areas that are pretty dramatic as well.What do you think, Rob?How do you react to that?

At some level, I think you're right.We're going to see the impact of changes and it's going to change the nature of work and what people do.And when I look at the decrease, what it tells me at some level is there's an opportunity to adjust what it is that we do with the fact that we don't have to do some of those sorts of things.So I can speak to this from an information systems perspective, maybe into the networking space a little bit better than some other places.But there is some of that work that is mundane and it is very much you had to technically know how to code, how to program, how to do some things that if you remove that burden and you free up people to do Other things, there's new opportunity for some really smart people to do some really cool things.And we need to figure out what that is.And I truly believe in human ingenuity and the ability for humans to adapt.And so I think what we're seeing in some of these numbers is where is there opportunity for some of these really smart people to do something different because they aren't spending their time doing some of the tasks that their expertise was necessary to get done?

Let's say that that's all accurate, because I think it is.I mean, we've been talking for years now, you and I, about the ability of AI to move humans up the stack to get us further away from those mundane technical details and more into the more rewarding tasks that humans are better at.But are companies going to hire for that?Do they need more of that?I'm not seeing it so far.Maybe.Yeah.

I've seen some projections that suggest organizations are beginning to recognize that if they aren't purposeful about hiring entry level positions, that they're not going to have anybody to fill in the positions that as people retire or grow out of them.So I think from a organizational management perspective, they're starting to recognize the threat to their ability to continue if they don't participate in the entry level job market.So I think what we're seeing at some level might be a pause as organizations are figuring those things out.

Yeah, I can see that.Although I'm still fairly concerned because of our current economic and political climate, if things get worse and either or both of those companies are going to tend to go into a protective mode and contract hiring and other spending to the extent they can.The other thing that was really interesting in this CES report that I think we're not talking enough about, we broadly, not just the two of us, is that there's also a reduction in earning in these same AI exposed professions.And so just to throw more numbers out, advertising in public relations is having a 3.2 percentage point employment shortfall.That's 3.2% below the average.But the earnings shortfall is 6%.And we're seeing that kind of across the board with these AI affected professions.That is data that doesn't show up in unemployment statistics.A lot of those, a lot of the positions are not college required.So it's not like they're going into a job that's kind of in the same field.Maybe instead of going into network design, they're going into hardware tech support.So it's not that it's that they're taking jobs that aren't really in their field and don't really require the education that is not good for us in higher ed, especially when it comes on the heels of some general skepticism about the value of a college degree.Rob, you're an administrator.What do you think?

Yeah, I think this is the big elephant in the room at some level, ensuring we are providing that value added, ensuring that when students leave having paid for a degree that they get meaningful work out of it that they wouldn't have otherwise.I do think society, whether it's politically or just people choosing where to send their students or asking that very question, I would argue, at least in the business community and in business colleges, this is why it's important for us to be engaging with those who are hiring our students, talking about what is it that they're seeing, what is it they want?Not necessarily tomorrow, because higher education doesn't pivot to the level of being able to do that tomorrow, but at a big picture, as we look at where we're going, what are the things that our students need to be prepared for so they can thrive as they get hired and demonstrate that there is that value from that degree?

Let me add something to that.I think we also need to build partnerships with those employers for experiential opportunities that may go beyond traditional internships and may include traditional internships.And one way that I think we need to talk to them is to say they have a responsibility in all of this.Technical skills are somewhat being commoditized, the kind of declarative knowledge, factual stuff that we're good at traditionally.If they want people that are going to be able to do the jobs of the future, those jobs are going to be highly contextualized, not just to an industry, but to a sub industry and sometimes to particular companies.And we can't do that.We can't provide that level of contextual knowledge or more importantly, tacit knowledge that you get by actually doing the thing.We can talk about how to do the thing and we can have them do the thing in small contained doses.But to really be out there and experience the messiness of being in a live organization, at best, we can simulate that.I think employers need to take on that responsibility.They haven't really.

They do in pockets.And I'm going to brag a little bit at Washington State University without necessarily naming names here, but this semester I've.

Just named a name.

But this semester we've had two different.We called them hackathons with an external company giving US real business problems, where they sent executives from their company to engage with our students as they were solving these real problems, giving them feedback about what they were facing and what they were doing.And it's been very well received.Even today, we have our board of advisor members for the MIS group here at WSU are doing an AI creative challenge with our students where they're bringing real problems they're facing and giving our students opportunities to answer them, engage them, and this is something they want to do.And setting these up and getting them excited to do it has been super easy, if you will.It hasn't been a begging exercise.It's been a how can we partner?How can we do things?And so in many ways, I think there is a willingness, given the right pathways, to bring externals into the conversation of helping our students be prepared.

Yeah.To be clear, there have been companies all along that have worked really hard to partner with universities, but that needs to expand.I think there's a lot of work ahead of us.We can't keep doing the same things we've always been doing.That's a recurring theme with this podcast, but I think we're getting forced into that more and more.I had a conversation with some faculty and doctoral students who were teaching yesterday, and the idea of how to do assessment, not assessment for accreditation, but to assign grades to the point where I'm kind of leaning towards, why do we have grades?But I'm sure that would go over well.But it's a lot of work.In an email you sent me the other day, you asked if we were at a messy place in the hype cycle.And that has to do with fatigue in a fairly direct way.So can you expand on what you were thinking there?

Yeah.So one of the things that I'm seeing in some ways is over the last two or three years, we've been pushing, what are we doing about AI?What are we doing about AI?And a lot of excitement, a lot of fear, a lot of oh, nos.And then we start doing some things and it's working or it's not working.But the question keeps coming up about, okay, now this new thing is happening.What are we doing about that?And I think there's some people who get excited about that.They like that change.But I'm seeing more people who want to, I guess, settle.Settle may not even be the right word, but lean into the things that they've been doing for a while to get that confidence of what they're doing.And almost, I'm afraid, people are Getting tired to keep fighting the good fight to ensure that we're assessing well and we're doing all these things well.Because once you've revamped something one year and you're asked to revamp it again the next year and then the next year, at some point it's why, why do I keep going?It feels like it's this new mundane thing that I have to do.So I am a little bit nervous that we're going to see some burnout from the continual change process that we're going through.

Yeah, even people that are happy to swim don't want to swim in turbulent waters against the tide.I think that's where we are.I'm getting a little bit of a sense where the resistance is dropping to AI.There's an inevitability that this is just the way the world is going to be and I need to figure out how I'm going to live in this world.But there's also a cooling off of the initial excitement.

That comes with the hype too.I think to get AI out the door.We were being promised sunshine and rainbows and unicorns and all these things.And as people started using it, it's like, okay, this does some nice things for what I'm doing, but it hasn't drastically changed how I do my job.In some ways, like, okay, I'm going to use it for these things where it's taking care of some stuff that I didn't love doing, but it's not putting us in a world where I'm sipping lemonade on my front porch all day long and the machines replaced everything that I'm doing.And then.So I think in some ways there's some disappointment in the over promise of what was possible and the under delivery, though I think in the long end of the tail of what AI does, we're going to start to see it reach those promises.But I don't think it's happening overnight like some of the commercials or some of the posts you saw on social media would have suggested.

I agree with all of that, but I think the nuance here is that AI's capabilities are just expanding.If you know how to use AI, we really have not gotten anywhere close to the point where it's just there.I mean, it is in some little things, but to do some of this knowledge work that we do, it still requires a certain skill level.And I think that's where I'm seeing a shift in the source of the resistance to AI.Early on, it was this true belief that AI is bad and there's still a fair amount of that out there.But now I'm seeing the resistance being more based on maybe a lack of self efficacy.I don't know if I can learn this stuff.The perceptions of a huge learning curve, fatigue or even general negativity about higher ed in society.It's just kind of hard to tell.Which makes it tough for people that want to encourage faculty and staff to be using AI.

Well, and the other thing, I think we're seeing some resistance of AI from a societal perspective of I no longer know what a person believes or thinks because everything that they're saying has been filtered through an AI lens to make them sound better.I read an article the other day about someone analyzed how much AI was being used in speeches given to Parliament in the UK and it's gone up to where a significant amount of the speeches that are being read are through AI detection tools.Like them or not, what they are, but they're saying based on everything they've looked at since 2022 when ChatGPT came out to now, a drastic increase in what AI was being used to say things publicly as part of people's political beliefs and there's some resistance to say, what does that mean for what my representative truly believes?Are we seeing those sorts of things in that public space?For me, it was kind of like what am I doing as a professor in some ways to where the students question the value of what I bring to the table.If everything they're seeing looks like it's coming to them through an AI lens that I put it through, does that begin for them to question what is the value of engaging with me, the professor who should be an expert in this topical area that I'm educating them in?So I think that's part of it too is authenticity and what people are really getting from the people that are the experts?

Yeah, I think that's once again this idea of leaning into our humanity, not just for students, but for the way we teach.If we have more of that connection and move away from some of this mass produced learning, I think maybe we're going to end up better off.Along those lines, I want to put a brief plug in for the system that our guest from last episode, Bill McCumber, talked about how he was using AI to help do more authentic assessment.He's actually getting ready to launch, so there'll be a link to his new system in the show notes.So you can go back and listen to episode 39 if you want more information on that it's AI goes to college.com 39 will get you there.Yeah, I think we keep saying it over and over again.We're just in for a weird time, kind of the long, strange trip.But there was another thing that you sent me that I think is pretty interesting around the recent vulnerability that kind of caused some havoc on the Internet.It's the PHP vulnerability.So maybe you can give us the ELI5 version of what that was.

Yeah, so there was.I came to work earlier this week and a lot of talk over, did you see this website was down, this website was down.And so there were websites down.And so as a security person and not to get into the weeds too much, part of me said there was a hack, there was a vulnerability, there was a thing that happened, it seems like it happens at least every semester, that something notable is taken down.And so what I learned throughout the day is that this technology called php, we don't need to necessarily talk about the acronym, but what is it?It is the software that queries databases behind the scenes in really any website that has input forms or output forms that that does the magic, that makes the data show up, that's being stored somewhere.And there was a vulnerability that had been in this code that had been known for years, but it had been so hard to get to and to exploit that they left it alone.It was almost like we were protected by difficulty.Well, AI, because of how good AI is getting it, ultimately found pathways to exploit this known vulnerability, giving people access to every server that was running PHP websites.About 70% of them use PHP, which meant much of the Internet and the websites we go to were affected by this.And so, thankfully, there was a patch that was pretty quickly able to be deployed.But the first step in doing things to protect everybody was take all of these websites down.And so that's exactly what happened.A bunch of websites came down, patches got applied, they came back up.We didn't see a lot of Internet downtime because this all happened pretty quickly.But for me, what it identified was two things.One is the quickness of AI that's going to be able to exploit things that affect our lives.As we take the benefits from the Internet, that take the benefit from being connected all the time with AI, that benefit may be taken away at some level as they try to navigate some things.So we're going to experience some weird in that.The other part is human decision making that's necessary from a world of protection and governance of these benefits and what we want to do.What we could get away with in ignorance at some level or in obscurity or whatever justification we have for things not being perfect or allowing imperfect things to be available.We're going to have to step back and say, as leaders in companies, as we're pushing things out there, what are we doing to put guardrails on our product, on the things that make us money as an organization to ensure that AI tools are not going to come and steal money from our employees or our customers or steal their information or whatever that looks like.I think the game of cat and mouse for connected devices is going to be filled with these sorts of news stories over the foreseeable future.

Yeah, it's kind of a weird yin and yang thing with AI where it was not only used to exploit the problem, I'm pretty sure it was used to fix the problem.So.Yeah, and I think we're going to keep seeing that and we're seeing that in higher ed.McCumber's system is a good example of that.AI makes it really easy to cheat.So the response is that Bill puts into place this AI based system that makes it harder to use AI inappropriately.I don't know what it all means yet, I don't think anybody does, but it's pretty interesting.And I think we need to be ready for continued turbulence as much as we'd like to avoid it.

Yeah.You talk about McCumber's system.I was in a conversation yesterday with some of the leaders here in our college, and at some level we're asking students to do certain tasks without AI because it's important for them to demonstrate what they've learned or whatnot.So we're not saying all AI all the time because that's probably not the right solution either.And so the idea of being careful with using AI tools to offload our jobs when we're asking students to write answers, not using AI to demonstrate what they know.And so that's that yin and yang part of it again as well is at what point is too much AI and what we're doing as educators detrimental to what we're trying to accomplish in building credibility with our students into doing those sorts of things.

And I think you have to go beyond simple transparency to explain why this enables me to give you a level of feedback that just isn't feasible with normal ways of evaluating things.I don't know.I think one of the big messages here to take it up a level is don't believe the hype or the anti hype.Anything else on this one, Rob before we move on.

No, I think I'm good.

All right, I want to take just a minute.On a personal note, I want to personally thank our now former president of Louisiana Tech University, Dr. Jim Henderson.You're going to find this hard to believe, Rob, but throughout his three years at Tech, I've never had a bad thing to say about him.He's just been fantastic.I want to publicly thank him for what he's done for Louisiana Tech University.It's going to be a hard act to follow.But now I want to get on to some practical applications of AI.Have you got anything cool and new to share?

I've been working with a colleague of mine who's set up his own servers that he's magically doing stuff where agents go off and create things.And in 30 minutes, he's creating websites that help students identify certifications that match the job roles that are being hired for.And just the other day he created one for marketing and showed me what it was and I said, that is awesome.Can we do this for mis?And he says, give me 30 minutes.And in 30 minutes he made a webpage that would have taken me a week or two when I was a developer to create super, super cool.What is possible as you begin pressing into this and being able to think beyond what is just a simple design that's really interesting.

And a couple of tools that I will recommend.I use one called Lovable.I think it's Lovable app that can create some pretty amazing websites or even micro apps, little apps that are quite useful.I've got one that manages all my research projects in kind of a Kanban board and then ChatGPT sites, and I think Claude Artifacts is kind of the same thing.I've created some really cool things with ChatGPT sites.They can be tutorials, they can do live updating.That's what we use to do.Our AI goes to college, you know, a little news briefing.It's pretty remarkable what you can do just through a chat interface.But I had a couple of different uses I wanted to mention.One is so trivial, but so useful.I have a spending account as part of my eminent Scholar chair.And keeping track of that is kind of a pain.So I ask our administrative person to just give me what I've spent over the last year.And so I went into the office yesterday and there was a printout in my mailbox with a table on it.I could have said, hey, could you put this in a spreadsheet for me?But she's got plenty of Stuff to do.I don't need to bug her.I gave it to Gemini, said turn this into a Google sheet, like two minutes it was a Google sheet.And then it kind of went beyond.This was really cool.And it did an analysis of it which I had not asked for.But I noticed there was a second worksheet tab at the bottom.I clicked on and it was here's what you spent on travel, here's what you spent on this, here's what you spent on that, which was pretty interesting.But if you have something you want to turn into, from a format you can't use to a format you can use, put it into your favorite AI tool and in all likelihood it'll change the format for you.The bigger thing I did recently is I put on an AI workshop, kind of an advanced level workshop for the university around bespoke instructional materials.

Those are fancy words, Craig.

Materials, bespoke.Just kidding.So yeah, that's back when I was kind of having to dress up all the time.You could get a bespoke suit which kind of fitted to you.So custom.That's not as much fun as bespoke smart ass.But, but like we have a class CIS 1253, that kind of the first year, everybody's got to take it heavy emphasis on Excel.Well, in, I don't know, 15 or 20 minutes, I took some materials from that class and said I want an Excel handbook.Don't give them the answers to anything, but I want them to have a handbook that walks them through the skills necessary to solve these assignments.Great.A little bit of editing and you're done.Now you've got this thing that's not just give me an Excel tutorial, it's give me one for this particular class where the problems are framed the same way, goes into the same level of depth, that sort of thing.And then at the other end of the scale, I've got a prompt that gives me a teaching guide for my doctoral seminars.And so our seminars are probably like yours.There's a half a dozen or so really dense articles that the students have to read and we sit down, we discuss them and go over them, that sort of thing.Well, it's really hard when it's the eighth time you've read these articles to think about what's going to be unclear, what might require some background that the students don't have, refer to some theory we haven't covered yet, that sort of thing.So I have a fairly extensive prompt that I give.I use codex, but you could use anything to do this with.I Give it my PDFs, give it this prompt, and it turns out this.It's usually about 15 pages.Here are the key points from these articles.Here's how they tie together.Here's what students might be confused about, all that sort of thing.Most of that I would have figured out on my own.But to have this thing created in just a few minutes at almost no cost to me is pretty phenomenal.And I could turn that into a reading guide for the students.I'm not sure I want to do that because I want them to struggle a little bit.But this general idea of creating custom or bespoke instructional materials is pretty powerful and it's not really hard to do.My prompt is long, but it's not anything that's really complicated.Have you tried anything like that yet, Rob?

Yeah, I do things like that quite regularly where I'll have a section of text, even for a class, and I want to make an active learning activity out of it.And I'll say, here's what we're talking about.Here are the key points.How do I create something that helps me to engage the students actively in the classroom around these topics and what I come up with some of the ideas that it gives me on my own, maybe, probably, but it does so in a way to where within 15 to 20 minutes, I have something that I'm like.I can walk into the classroom and do something that is creative, engaging, and fun in ways that I probably would not have done that as often or as frequently as I did.And it really makes it a lot easier to move from sage on the stage, the stand, and deliver a PowerPoint lecture to the engaging guide by the side approach to learning that I think is going to be what we need to be doing to let students experience the messiness of business while they're in the classroom.

And not just business, but any field, really.

Absolutely.

Well, Rob, since you have brought things full circle with that comment, I think it's probably a good time to close.Do you have anything else?

No, Craig, I'm excited to see what the next episode brings because I'm sure there will be something new and exciting that drops in our lap between now and then, as it seems like that's the world we live in.

It is indeed.All right, that's it.Talk to you next time.Thank you.