Welcome to another episode of AI Goes to College, the podcast that helps you higher ed professionals figure out 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 of Washington State University.

Joining us today is my friend and colleague, Dr. Tom Stafford.

He's professor and chair of the Management Information Systems Department at the University of Memphis.

He returned to Memphis after, I don't know, what, 10 years or so at Louisiana Tech?

I think nine years, something like that.

Close, close.

He was the J.E.

Barnes professor of Information Systems and eminent scholar and also the doctoral program coordinator.

But Louisiana Tech's loss was University of Memphis's gain.

Tom, we're happy to have you on AI Goes to College.

It's good to be here.

I remember you two doing your podcast live in Nashville at the conference and I was impressed.

The room was full, the interaction looked interesting, and I had somewhere else I had to go, so I stepped back out after I peeked my head in.

But you guys have a very interesting invisible thing going on here with this podcast.

That's because there was free food in a bar.

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

We invited Tom on because not only is he the co host of a podcast that I sometimes do called Cyberways, which looks at cyber security, but he also is an expert on programs that are integrating AI into the curriculum, both from the perspective of current curriculum, but also new programs.

And we've invited him here today to talk about new programs.

So broadly speaking, we're seeing schools do one of three or four approaches to integrating AI.

One is AI literacy, whatever that means, and we can talk about that a little bit later.

The other is let's have an AI course or two somewhere.

The third level is let's have some sort of AI certificate or minor.

And I would lump in maybe co curricular micro credentials into that bucket as well.

And then full board degree programs in AI, whether they're majors within an existing degree or new degrees.

And Tom, you've been working on a new degree program?

Actually a couple of new degree programs.

Right.

We're going live with the graduate certificate in fall and we just got approved in the catalog for a concentration in our master's degree that uses the same set of courses, plus 1 for 4 course graduate certificate and then adding a cyber ethics class to it makes it a five course concentration in the master's program.

We have two classes enrolled for fall in the certificate and it's going to be interesting to see how that goes prompt Engineering and AI assisted application development are what we're starting with.

We'll follow up in spring with AI Assisted database and AI Assisted Analytics.

The approach has largely been AI assisted.

My sense of it has been we're not equipped in a business school to help anybody become an AI wizard and build algorithms and design platforms and all that stuff like my kid is learning how to do down at Texas right now.

He's a propeller head.

He's a programmer.

He wants to build systems.

But I'm more into having my students ready to use AI in the workforce because we're headquarters city here and everybody's requiring their people to start using it.

What can I say?

You know how I feel about AI.

I try to.

It's not that I try to eschew it.

I'm trying to remain a, an objective observer, which means I don't use it very much because I feel like it'll color the research I'm doing with it.

And my provost tasked me with getting a sense for how the university community views AI.

And I'd shared with you some of my, my early analysis on that.

The, the power user analysis is what I'm thinking of it as, but I've also got a very clear sense for, for how the haters think about it.

I call them the haters because that's, that's the construct.

Well, let me interrupt you.

You've given us 15 rabbit trails there all in one in here.

And I want, Yeah, I want to.

Dive into your program a little bit because I think that's interesting.

And so can you describe that for us?

You told us some of the classes you were doing and the MIS hat on me sounded like they were MIS classes that you repurposed with AI to be more AI focused than what an information systems class would have been otherwise.

Is that assessment of the direction you went?

It is fair, though I wouldn't admit it to my competitors in the programs across campus.

I'll think of it as AI literacy, broadly, broadly considered.

We want the managers of the future coming out of our program to be able to integrate AI into their work.

And that's going to mean being able to, I think, largely do analysis that drives decision making with AI.

So the analytics side of it's going to be quite interesting.

We're a SAS campus and we're using SAS Viya, which is a copilot integration.

And it's been popular for a year or two now and going to be even more popular, I think, as students take to it, because it's easy to use you know how difficult SAS is as a standalone statistical application, but using the AI integrated version of it is, I don't want to say child's play, but it's way, way easier than SAS programming.

So as you put this together, Tom, one of the things that I struggle with as an administrator myself with putting programs together is staffing programs in a world where resources are limited.

Were you able to do this in such a way that you could redeploy resources you already had, or did you have opportunity to bring in additional people to do something like this?

What did that look like from a deployment standpoint?

Well, we were lucky that one of our recent hires, the other half of our doctoral coordinator, Hui Gang Liang, his other half, Ya Zhongshui, we call her lucky.

She was already doing AI assisted analytics down in North Carolina and is going to basically port that right in.

As our, our fourth applied AI class.

We had a couple of people on ground already using SAS Viya in the master's capstone program.

And we have an artificial intelligence class and a machine learning class already as part of the master's program.

But it's more wonky than the applied AI approach.

I'm, I'm seeing this as more AI for the.

I don't want to say the common man, but I don't have a better term in mind right now.

A gender neutral term doesn't come readily.

The common person, if you will.

AI for everybody.

In fact, we have an undergraduate minor called AI for all.

And that's largely the sense of what I'm trying to do at the graduate level is make sure everybody comes out of this program able to use whatever AI technology their employer wants them to utilize in the pursuit of business productivity.

What I heard you say you had some courses already on the books that you were able to utilize as part of this.

And then you did have a person coming in that brought some new capabilities.

So there's a little bit of resources there to help with that.

The courses in the master's program remain.

So anybody who takes our msis concentration in applied AI will take the applied stuff on top of the heavier stuff, the artificial intelligence.

Machine learning one, machine learning two.

They'll be taking that in addition.

So I'm thinking of this as a more user friendly approach to.

I'm going to go for AI literacy since I have Craig's model here in front of me.

Or maybe this is your model.

This model?

Yeah.

I was playing around with a three level model of literacy.

Literacy strikes me as a good way to describe what we're doing because the folks who are coming in will get this early.

We're hoping we use our graduate certificate as an on ramp to the master's program.

You can enroll in it with an easy application right out of your baccalaureate program.

So it's like a two tiered approach to enhancing our enrollment.

And like most universities everywhere, we were heavily invested in the international sector for our graduate program and are having to learn how to do without that.

And so this is intended to apply to locals who want to find a quick and interesting way into graduate school and who at the same time are sensing the wave of the future coming at their workplace.

Because we train a lot of people who work their way through their master's program.

They're sensing the need to school up on AI at work here at the headquarters companies in Memphis because everybody's using it.

So let's see if we can take a big step back here and put a little structure around all of this to generalize it.

So why did you think about AI specific degree or certificate programs rather than infusing AI throughout the curriculum, either in information systems or in business more broadly?

Well, that's a competitive thing.

I guess I'll get hollered at by the boss if that being the folks up the food chain.

But when I came and interviewed, I already had learned that there were other initiatives aiming towards curricular development in AI.

And it seemed to be, I don't want to say flavor of the month, but it seemed to be a very popular alternative to traditional academic programming.

So being the former marketer I was, I fully intended to come in and be first to market, get first mover advantage.

And I proposed through the state curricular proposal system right off of the bat.

Got approved before the other programs did and am already enrolled before the other programs are, except for the undergraduate minor of course, that had already started and we were already providing programming for that.

Not as much as the other departments in the college, which I found curious.

The management department here has got a whole lot going on at the undergraduate level for the AI minor and the baccalaureate degree.

But my clear sense of it was the folks at International Paper, FedEx, Autozone and name your flavor of the month for the large companies here in Memphis already leveraging AI.

We've got an XAI footprint on the ground here and we just felt that people who knew AI had a better chance in the job market here for locals who are training to enhance their employability in the local job market.

I think a lot of schools are struggling with this.

Should we do AI across the curriculum or should we do AI specific programs?

Yeah, that's a tough.

Let me put voice to my bias.

AI throughout the curriculum is a technology thing that goes beyond our capabilities as a business school.

Our job is to help people run a business well using the available technological tools, whereas, say, in engineering or computer science, their job is to help people understand AI well enough to build it, manage it, maintain it.

But.

But isn't exactly what you said what business schools could be doing.

So we don't have to get them to get our students to build the AI tools, but they should know how to use them in accounting or finance or whatever else.

Don't you think you hit the sweet spot with accounting?

You know, my research revolves heavily around auditing and audit automation, and I follow Miklos Vasarehely up at Rutgers very closely, and he's been on the record now for a couple of three years with his star doctoral student, Abby Zhang, saying that basically auditing is going to be a profession in which the accountant supervises in live time automation run by AI in accounting.

That really tweaked my interest several years ago, and I've been doing work in that area.

What I want to get a sense of, though, because people are struggling with this and you're way ahead of the curve here, is I want to give our listeners some insights on what your thinking was.

So.

And I'm going to Rob, feel free to jump in here and Tom, tell me if I'm off base.

What I'm hearing is that you wanted to have a marquee program that could draw people into the school, which is probably not as straightforward to do if you're going to put AI across the curriculum.

We're putting AI across the curriculum is easy to say.

It's hard to communicate exactly how in a succinct way.

We have this MBA concentration in AI that takes a business focus.

That's something you can make a pithy slogan around that will bring people into the university and into the college.

Is that something that drove your thinking?

The short answer is yes, it was.

It was pragmatically an enrollment strategy.

We lost tons of students in the master's program when the visa issues hit.

And my brief coming in was to find a way to engineer programs that would bring more local engagement in.

And AI seemed to be the best, quickest and most appealing way to do that.

It was an enrollment strategy?

Yes.

We were trying to maintain viability by offering topics that we think that the local community is interested in taking.

So, Tom, that seems like A great short term strategy and I can see that working to solve this immediate need.

One of the things that I've struggled with is if I think back of history and the E commerce programs that got built up, eventually what we taught in an E commerce program became part of almost everything we do in business because almost all businesses run through websites and E commerce sorts of things.

Five, 10 years down the road where AI is not something unique and different, but it's part of everything we do.

How do you see this evolving?

Have you looked in your crystal ball a little bit to see what the 5, 10 year pathway looks like?

I originally came here to Memphis as a junior assistant professor as the E commerce expert.

I taught classes in E commerce.

I brought marketing sensibilities having been terminally degreed in that area.

And very quickly before I got to tenure, it was exactly as you described.

It became mundane and infused.

It was not E commerce anymore.

It was simply the way business is done.

I think that that is the, the medium term outcome for, for AI.

But then I've got a five year contract and then I retire.

My goal is to get enrollment up and keep the program viable.

We're operating under an economic model here where we die on the vine if we don't generate graduates.

And the quick way to generate graduates is take advantage of the AI wave.

I'm shamelessly jumping onto that bandwagon.

I want to give a counterpoint though.

I've been harping the we should look to what happened with E commerce to understand AI for a long time.

And I think that's mostly right.

Where I think it's different is that after a period of development, E commerce kind of spread and what we applied it to and in relatively incremental ways.

It didn't have these big leaps where, oh, this is a completely different thing.

Like, okay, so we started out with websites and fairly static things and then it became more dynamic and then we had new intermediaries that came in and we could all name these various little things that happened.

But with AI, if you look at a chatbot versus agentic AI, it's a different thing.

And I understand where you're coming from, Rob, and I'm not sure I disagree, but I'm not thoroughly convinced of, oh yeah, this is just going to be the way we do things.

I think that technological frontier, that horizon, is going to keep getting pushed in big ways.

So I don't know, it'll be interesting to see.

You have a very valid point.

Having seen the E commerce example and lived it myself as as an element of it.

Some of that's going to be here.

We're going to diffuse throughout the curriculum with it.

But some of it is different.

This is a discontinuous innovation, and it's discontinuous in ways that we can't yet see the outcomes of.

So, you know, I'm a science fiction buff and I think back to William Gibson and Kurzweil and all of that.

There are things that are going to happen that we don't currently expect but might prognosticate on.

They don't always portend to be good things.

I'm a little bit concerned about, God forgive me, the future of education.

I don't know that we humans have that much currency left in terms of being able to deliver knowledge in context, like we're currently in electronic context, as compared to agents being able to do it in our stead.

I know how these things work.

Having studied distance education and the readiness with which university administration hopped onto the bandwagon of let's.

Let's make it cheap, easy and quick and go online, I can see them doing much the same sort of thing, possibly with AI.

I hope not.

But it's.

Some people are going to be financially pragmatic about it and give it a shot.

The other question is what happens to the job market where we're still trying to figure out where our graduates are going to go in a world where some of the jobs they would typically be schooling for aren't jobs that will either exist or be economical for humans to do as compared to AI?

This is an example of my own son Joseph, who was a.

A real good programmer at Epic Systems, the medical systems provider up in Wisconsin.

C Sharp programmer.

Two years in and part of it was burnout, but the other part of it was he realized his job was about to be obsolesced by AI and so he decided to, like Pogo used to say in the comics, I have met the enemy and he is us.

He went back to school to take a master's in AI and he's going to go take a doctorate in it, probably at U Dub.

At least that's where he'd like to go.

He really likes Washington.

I'm worried about the job market and I'm worried about it because I just don't know what happens to education when the job market changes.

Yeah, that's definitely a serious concern, but I want to go back to this idea of AI literacy.

So AI programs have their place.

We're working on one here that I hope will one Day be approved.

They have their place, but they're not for everybody.

They just aren't.

I'm not arguing against AI specific programs, but I think they need to not be positioned as an either or against AI literacy, which I'm not a fan of, the across the curriculum model because it never seems to work.

But ways of building AI literacy across our entire student bodies, I think that's a different thing than what we've been talking about.

And I'm going to make the argument that we're seeing a fundamental shift in how knowledge work gets done.

Oh yeah, that's what makes this very different from E commerce.

E commerce was about transactions and smoothing information flows and that sort of thing.

But we kind of did the work the same way we've always done the work, maybe a little bit faster.

Maybe you didn't need somebody to process purchase orders because they were done online, but knowledge work was still knowledge work.

The knowledge work of the future is not going to be the knowledge work of the past.

And this is where I think a lot of universities that are foot dragging are doing a disservice to their students.

We have to help them understand how to engage with AI in ways that elevate their humanity and their human contribution to knowledge work.

And that's what I think we're not doing very well right now.

So tell me where I'm wrong.

Rob and I were talking about that before you joined, before we started the roll in the tape.

My notion of AI is that people coming out of a business school are still going to be decision makers regardless of whatever decisions they're engaged with.

And AI is a very handy tool to enhance decision making prowess.

It, it allows you to gather data, organize data, make sense of data.

But I don't know that AI is going to be that handy in terms of deciding what to do with the data.

Because human intuition is typically where the decision making, the good decision making comes in.

The, the creative insight that leads to inspired decisions as compared to routine decisions, let's call them automated decisions, is still the distinction between the technology and the person.

I go back to decision support systems and Gori and Scott Morton.

The whole notion of deciding what to automate and what to lead to human judgment.

I think we're seeing that model writ large as we begin to decide how AI enhances human decision making capabilities as opposed to my scary notion of how AI replaces our value as an element in the business process.

So where I see that is an interesting argument, Tom, is from the perspective of the vast number of faculty that are engaging with preparing our students.

And we have pockets of people who are pushing the envelope, who are able to do exactly what you're talking about and really push about what's possible.

Where does it help and how do we make the decisions?

And then we have some that completely resist doing anything with it and spread across that spectrum.

And so one of the challenges I see that I don't have the perfect solution for yet is how do we ensure that throughout the curriculum we are preparing students for a world that has rapidly changed expectations of what a student coming out of our programs looks like?

I don't have the answer because we don't see far enough forward to know what's going to be.

I think to default to Craig's point of view, our best course is to make sure they know how to use the tools so they can apply the tools.

They see the places it could be applied opportunistically.

I would pivot to another point, though.

In academics.

Part of my study with use in the campus community has revolved around the predictable concern about students using it to get one over on the professor.

And Craig knows that.

I decided a long time ago I wouldn't worry about trying to catch people doing it.

I'd force them to do it.

The whole notion of you're writing a term paper in my class and you're going to produce it with an LLM, and if it's not perfect, then you've not done what I've expected you to do.

Students have to be dragged to that a little bit because many of them expected to be able to do work that was not required to be done by AI.

With AI, everybody's trying to catch people cheating, so to speak.

I don't see it as cheating so much.

I've been very pragmatic about academic dishonesty.

My.

My typical philosophy is people learn regardless of how they engage with the information, Whether they do it honestly or dishonestly.

It's still information processing in the noggin, and that's cognitive learning.

They just.

I think they learn better and also more wholesomely if they do it an honest way.

Which is why I require my students to write a term paper with AI.

And that's another part of my study for the provost.

I haven't generated models out of that, but it's very easy to see because it's qualitative research and you anecdotally discern what they're saying.

Students find it to be a very interesting challenge.

They see very quickly the advantages it gives them in terms of being able to organize content, search for information supporting their thesis point, and to make it look nice, to dress it up, the authorship part of it.

AI is editor is the construct I see emerging there.

They also very readily say, this is way more work than I expected.

So I often liken using AI in courses to the old notion of preparing an online course.

It's a time and a half job.

So if you're going to use AI to create your term paper, it's going to be more work than doing it yourself.

But the advantage is you'll have AI experience that you can then port into an executive position where you might write memos with AI, draft emails with AI.

My dean is big on this.

He, he always writes a lot of emails that could be politically sensitive because he's the guy who gets to deal with the hires up and all the political issues.

And so he uses AI to vet the propriety of his messages before he sends them.

He's a big proponent of that.

We talk about it in faculty meetings.

Sound familiar, Rob?

All right, so come on.

You're one professor that your students get to see, and I love your approach.

I think it's one of those definite things students need to learn.

I would put you in that first category.

I mentioned the person who's embracing it, who's figuring out the ways to do things very intentionally.

How does that ability, that way of thinking, how does that flow through to others that aren't necessarily as quick to jump on board?

And I'll even move out of the realm of information systems.

You're working a college of business.

You may very well have people in finance or accounting that say you still have to do math.

Math is the thing that's important.

Where does this come in as you engage in conversations across disciplines?

My sense of it in my analysis is the.

The haters, and I'll call them that unapologetically, because in the humanities, they figure it's not AI's job to create art or to enhance thought.

It's what they hired me to do and train people to do.

But as they interact with people like me and people like in the engineering department, where they're building AI programs unabashedly oriented towards preparing students for the coming future, they learn from them whether they like it or not.

So the people who are right now saying, well, AI can't make good art, are talking to the people who say, gee, I love to use Claude because it makes my programming much, much better.

And when they have lunch with them and talk about that stuff, inevitably some Cross pollination takes place.

So I, I have a construct in my model that.

Let's see, it's like Craig's grassroots.

I'm looking at it on the wall here.

Craig's grassroots notion, workplace diffusion of influence mediates by between or actually as a moderator between, influence transmission for the people who are adopting it and the people who are avoiding it.

I think eventually everybody gets on the bandwagon because they see the people who are really enthusiastic about it, having fun with it and they see their students benefiting from it.

And sooner or later it becomes something that you can't avoid doing because it's working.

So let's take a beat here and pursue something that you just said.

Okay?

I'm a big fan of, of not worrying about the haters.

They'll come along.

They won't come along.

This is another reason I'm not a big fan of anything across the curriculum.

If somebody doesn't want to use AI in their course, don't make them use AI in their course.

That's just a losing proposition to try to force them.

But what I'm seeing some, some issues with that are getting a little bit better is a real lack of coordinated faculty support.

Support for those who want to engage with AI.

But maybe they're not plugged in.

They're not their social network, their professional network doesn't have people like us who are really into this and pushing the edge on it.

They don't know where to begin.

I'm starting to see schools being a little bit more purposeful and going beyond.

Let's have a workshop.

The workshops are great, but they're one offs.

They really don't carry their half.

Life is terrible typically.

So like we're getting ready to do a couple of mastermind groups here at Tech where we'll meet periodically and hopefully we'll bring more people into the fold as we go along through the mechanisms you were talking about, Tom.

So I think that's something that schools, we've harped on this.

Schools really need to be supporting faculty that are willing to at least be open minded about this.

Forget about the haters.

I think right now it's a losing effort.

I don't want to say it's a waste of time, but I think it's resources poorly spent to try to drag them kicking and screaming.

Yeah, and what you mentioned there, Craig, is something that I think, to Tom's point earlier about preparing students with use cases for how they're going to utilize these tools differently in how they critically think and evaluate things.

The same, I think, can be said for faculty learning the use cases from each other, and I would go so far as to include staff in that conversation as well, is nobody has the silver bullet that says, here's exactly how you're going to use it.

What's more important is getting people who are using it sharing what it's capable of and how they're finding ways to use it and opening doors for those conversations so that way those light bulb moments can come on where people can say, aha, that's how I can use it as well.

And how do we tear down the, I want to say, the silos that we live in, where we aren't really sharing how we're going about creative new ways of doing our jobs?

I want to coin a little bit of a phrase here.

I think we need to set the context for discovery.

I think that's the focus we ought to have, is let's give people enough kind of foundational understanding of what these tools can do and then set things in place where they can explore, they can learn from each other organically, they can push the boundaries, maybe in some safe ways.

I think that's the way to approach this, rather than workshops or anything across the curriculum and just say, okay, look, we're going to help you understand kind of what's going on with these tools.

We're going to give you some examples, we're going to have discussions about how you might use it, and we're going to give you the tools to play.

Like we're trying to explore giving everybody that that's involved in the Mastermind paid accounts to ChatGPT or Claude.

So here we're going to fund you for the next three months, six months or whatever.

We'll talk about FERPA and all those kinds of cautions, but go play with it.

Some things are going to work and some things aren't going to work.

But that's what schools need to be doing, is setting the context for this organic discovery, which is what you all have been saying.

That's grassroots diffusion writ large, isn't it?

Yeah.

Yeah.

I think the points that you've all been making about having the tools available for people to experiment with resonate hugely with me.

It flows right along with my own naive approach to this, which is require the students to do their term papers with an LLM.

And I'm analyzing that data and I'm finding that they appreciated the opportunity.

Students are all about, gee, the professor's trying to legislate my use of technology with the folks who are trying to catch them using it, as opposed to, gee, the professor is requiring me to learn how to use this because he knows I'm going to have to do it when I graduate in another year.

I had a very beneficial experience both at Tech and here, watching the students learn how to use the agent to produce their written work.

And of all the user communities at my university and also at Tech, where I was experimenting with this, that I've analyzed, the students have the brightest outlook on this.

It's the colleagues on staff and faculty.

So some of whom are worried about the, let's call it the terminator factor, if you will, AI taking over the world and stuff like that.

So the conspiracy theorists are among the faculty and staff.

The students are very open to this, and I'm encouraged by that because that's where all of this has to be deployed.

They're the ones who are preparing to use this in the workforce.

Tom, what you say there, I think, is consistent with what I've heard from some of the industry people I've interacted with here at Washington State University, that they're excited to see what the students have been able to figure out.

Looking at our students as these AI natives, where they've actually, through their education, we're going to be graduating students pretty soon that have had AI available for them through their entire collegiate career.

And giving them that latitude in the classroom space to be able to explore and to play and experiment with the permission and guidance really does set them up to how do they solve business problems in the workforce in creative and different ways, because, you know, of the.

Let's call it the laboratory of higher education, where they're able to actually, you know, in your class, do exactly what you described.

But then they might take another class from Craig, who encourages them to experiment and play in some different directions.

And collectively, they're going to have a foundation as they discover and utilize these tools in different ways.

That's going to make the unthinkable possible, I would imagine.

You said something there, permission and guidance.

I like the sound of that, because that's the positive way forward.

Instead of saying, the sky is falling and AI is taking over the world, I think it's just like with any other new technology that comes out.

We have to learn how to use it because it's here.

It's not going away, genie's out of the bottle.

We have to facilitate.

Facilitate their understanding and use of it.

I believe both of you focused on students, which is certainly important, but I want to see what you all Think about this for faculty and staff.

So when we were talking about it initially, it seemed like there's a lot of difficulty in getting faculty and staff not just on board, but the experience with AI in a guided, structured way.

So Rob, let's take off your co host hat for a minute and put on your administrator hat.

And what do you think?

What are you seeing here?

Yeah, I think that's absolutely correct.

And I want to share an anecdotal story of something I just did here at Washington State University.

And I've done this with staff at both with the college event and your with the Voiland College of Engineering and Architecture and then also with the Carson College of Business staff in my own college.

And it was an AI workshop for staff.

They want to know how can I use Copilot?

Because that's our tool that we use here.

How can I use it to do my job?

And as a faculty member, who am I to tell an administrative assistant exactly how they can use this technology to do their job?

And so some ways I'm not sure what they do all day long as far as their work role.

And so I wandered around the building and gathered a handful of examples and had people show me what they did.

And as part of the workshop I spent the first bit, first half hour or so showing what is possible and then released the people in the room to begin trying to do something.

So identify roles that had friction that they didn't enjoy as part of their job.

How do you then take what you've seen and do something with it?

If you get stuck, you can't figure something out, talk to the people at your table building this community of people working together.

And then when we were done, there was a share out where people could explain what they did.

Again, furthering the community of hey, I love what you did there, we need to talk so I can see what you're doing.

And on the backside what we're going to do be doing is creating some teams channels or some other pathways of continuing to open the door of community.

So as people are learning things, they can help each other.

When people get stuck trying to problem solve in their use cases, they can work together.

And so it very much is, yes, we've been talking about this a lot.

The real way we're going to find out how to do this is to encourage you to, because we've shown you a little bit of what's possible, but then put you in a space where you can encourage each other and help each other figure out these use cases.

That's a great example.

But one of the things I want us to talk about a little bit is how to systematize this.

So how do we put this into something more coherent and something that might get to the point where it could sustain itself?

So what do you think the components are?

Centers for Instructional development.

You were telling me about Centers for Instructional development.

You were telling me about Professor Jacob at Tech and the program she's putting together, which sounds exactly like what our provost is putting together here.

What was the technology center that supported the online learning deployment is now becoming a merged technology center that supports the rollout of AI literacy amongst the faculty users.

I think our provost has seen what you have seen, which is the students are learning with or without us.

The faculty need to be enabled.

To enable the students in terms of AI literacy, the faculty have to catch up a little bit.

And so we're institutionalizing the process of.

I'll call it a safe sandbox.

A place where people can come and learn their limits and learn the opportunities, I suppose.

And I'm looking forward to that myself because I'm in the middle of enrollment management right now.

I'm having to use Power bi and I'm basically.

Basically having to prompt copilot to get the analysis I need, which is fine.

It's actually working quite well for me because I'm not a usual Power BI user, except he's a little slow.

And I'm getting used to that because it's what do you want for free?

Basically, I'm getting what I'm paying for in the Microsoft Office implementation there.

So a couple of things.

You mentioned the foundation of AI literacy.

Now, exactly what AI literacy is, is a completely different conversation that we can have another day.

Well, let's just make it a thing right now.

You also talked about room to experiment.

I know that you used that exact term.

But where we can have this kind of sandbox, safe space for experimentation so people can learn.

You learn the limits by exceeding those limits and then suffering some consequences.

The classic case of a kid touches something hot and learns not to touch hot stuff anymore.

But you've got to do that in a safe way.

Otherwise people become paralyzed and they're afraid to try anything.

I think one more thing that we might have seen come through so far is the idea of community, which I think is really the sustaining feature of all of this.

If you put people together that are all trying to explore, they can share and cross pollinate and learn from each other, we're doing that.

I'm working with A friend of mine, Brad Miller, who is a fellow podcaster on something called AI Assisted Author, where we help content creators take their content and shape it into a book that they write with the assistance of AI.

Not one that, hey, AI, write this book for me.

And the most valuable part of that so far has been the ongoing community.

Matter of fact, we're meeting tomorrow without any real agenda to speak of.

Just here's an opportunity to share what we've been doing, what's worked, what hasn't worked.

Are either of your universities going very far and creating some sort of a community of practice around AI.

We're taking baby steps.

The provost is determined, coming from Purdue where they did this, he's determined to enact the center for Instructional Development around AI here.

And that's where I'm going to be able to learn.

I'm able to learn from other people who have already learned.

In my own analysis, I've got that grassroots notion we talked about earlier where the, the power users and the haters can interact and share information and everybody comes to a center point where everybody feels comfortable with it.

An interesting anecdote.

I'm an AI neophyte and when I'm in the company, you guys, I feel like I'm, I'm out of my depth.

I don't know the terms, I don't know the tools, and hesitant to admit it.

But in a community of practice approach where people can interact with each other safely, share anecdotes, diffuse innovation, that's going to be an effective way for AI to become more user friendly, friendly in the faculty and staff group.

I think students are going ahead full steam.

They'll leave us behind if we don't hurry up and learn how to stay with them, is my sense.

We just make up the terms, so don't feel bad about that.

Craig.

What I think is an interesting, I guess, tension here is that some people, faculty, staff, look for administration to create the structures to teach them what they need to know to do their job.

And when something is as new as AI is, what I've seen is actually working better is when it's more grassroots in nature in the sense that it bubbles up from those who are doing things.

And we create a space really where people can show up and share and be a part of those conversations.

When I've seen meetings, workshops, whatever they are held, where it's promises of we're going to teach you everything you need to know, people walk away disappointed because there was that one thing they really wanted to know and nobody answered that question for them.

And so in this space of a brand new technology that has a whole host of possibilities when that space is created to bring people together who can wrestle with the messiness of learning that in a space together, I think that's where we're seeing some really cool applications getting shared out.

And so that's what I've seen work really well.

I have a counterpoint, so I don't disagree with anything that you all have said, but my concern would be that the organic or grassroots approach is not going to spread far enough.

And so it may be self sustaining around a small group of people, but something needs to push it outside of that circle.

And Tom, you could go into the whole strength of weak ties thing for us because you're an expert in that area, but it will stay within that group.

And so where I see administration's role is to set up the structures that will allow it to go outside that initial organic grassroots group to spread.

But that's a great way to get started, is to recognize, okay, here's a group that's already doing what we want to do.

Now how can we spread that?

The other thing that I think administrative involvement can do is signal that it's okay.

This is acceptable use of your time.

We're willing to accept some failure as you go through the experimentation process.

And leaving groups on their own may not signal that to the more reluctant adopters of AI.

But I could be wrong.

I have a thought in that direction.

It struck me for a couple of years now that universities mastering AI is a lot like in the past couple of decades universities mastering distance education.

It's a time and a half task.

It's more work to learn how to use AI to do your work than it is to simply do your work with up the extra time it takes to learn how to integrate AI into it.

That's how it's been with me.

It's not that I don't want to learn AI.

I don't have time to stop and learn it.

I do.

But you see what I'm saying in that direction, it is a little more work than trying to do it without it.

And if the university is willing to facilitate that, as they did with online implementations, it would give people course release time and financial considerations for having to spend the extra time doing online implementations of courses.

If they can do the same with mastering AI, I think you'll see a lot more uptake.

I also see two parallel issues, two parallel channels of influence, if you will.

The administration has the power to fund it has the power to mandate it if necessary, has the power to organize a learning structure for it, which I think is very positive.

And then the grassroots approach is where the credibility of the knowledge exists.

So you guys are administrators, you know that the faculty takes what you say with a grain of salt because your job is to manage them.

And they want to be independent contractors as faculty are.

If we have the communities of practice that have the grassroots power, where the haters and the power users can interact and learn, AI is not so bad after all.

Combined with university formal support in learning structures and funding structures, I think you get a very positive outcome.

Yeah, I would agree.

I want to touch on something that you said, Tom, that harkens back to a paper that we wrote 20 something years ago on grassroots diffusion and that's identifying people that truly enjoy using this new technology, AI in this case.

I love playing around with this stuff.

Matter of fact, I get up very early and a lot of what I do the first hour or two of the morning is play with AI because I like it.

So it's not a burden for me.

But what universities need to start doing is identifying people that enjoy this AI experimentation and tap into them as opinion leaders who can then spread some of these uses and help others learn, connect.

Them to the doubters.

Yeah, and that's the, that's the center for Instructional Development notion right there.

The what Rob was speaking to earlier form communities of practice that the university sanctions and supports and people will learn on their own in that kind of context.

I'm going to be a little bit of a skeptic here, and this is not related to anything in my direct personal experience, but a lot of times these centers for instructional excellence, a lot of times they're good at one offs.

It's a lot harder to build that community or if they build a community, it's very limited and they have limited resources.

A lot of times it's a part time job for somebody.

So I don't mean to be critical, but I think within, within colleges, within departments, you could do something kind of outside of that more formal structure.

Rob?

Yep, that's, that's what I'd agree with.

As, as you were describing the challenges with the centers.

Oftentimes what I see is the same people gravitate towards involvement with centers.

And the thing with, as I think about what's possible in a college of business and I've interacted with the college of Engineering, I've interacted with some people in ag colleges and so forth, is in these communities there's even sub communities and different people that you just need to get in the same room.

What I like to say is to break down the silos of our office doors.

Oftentimes we close the doors and we try to solve these problems themselves.

And so whatever it takes, whether it's within a department, within a college, within similar roles, staff roles inside of a college, that those are where I see the communities really helping each other.

And you've got six people that all somewhat do the same job.

And two or three of those people are seen being super productive because they figured it out that information is going to start sharing through those networks if we just make it easy for them to share, maybe communicate expectations that we want those people who have been those first movers, those people like Craig who just are getting excited about what is possible if they're allowed to take time to even sit down and share with their, their colleagues who haven't gotten there yet.

And that becomes part of job role a little bit, it's going to begin to bubble up more confidence in what can be done and more people engaging with these tools.

You said some something earlier in this colloquy that resonated with me, learning the use cases from each other was the quotation.

It's powerful.

It captures the essence of how we're going to succeed with this.

If I can get the arts and sciences and humanities people talking to the folks in engineering in the college of business, everybody's going to benefit from learning from each other if they stay siloed.

As you were speaking to, some people aren't going to learn a dang thing that they can use and other people are going to have a lot of knowledge that they can, that they could share, but don't have a mechanism to share it with.

And we won't be as efficient in the academy as we could be with our AI implementations.

Breaking down walls of communication is the most important thing.

Another phrase I'll share with you that I use when I talk about this is none of us should be inventing the same wheel all by ourselves.

That we should be sharing what we're doing because otherwise we're duplicating each other's.

Work for some reason.

Those old spinner wheels from back in the 90s popped into my head when you said that.

I'm a little reluctant to bring this up because it might be a rabbit hole, but one of the things that I worry about with staff is that as they experiment and gain efficiencies through AI, they're just going to get more Work piled on them.

I mean, with faculty, if I can make my grading more efficient or pick a thing, make it more efficient, we're, as Tom said, we're kind of independent contractors.

And so nobody's going to know that we're more efficient and try to pile more stuff on us.

But with the staff, this is a huge problem, not just in higher ed, but oh yeah, do all this stuff, take this risk, do this extra work to gain these efficiencies.

Oh, thanks.

Here's more work.

The business process re engineering era had the same effect with, you know, AI not being in the equation.

You made a valid point.

That's, that's happening.

I feel it in my bones.

Yeah.

And what I think is interesting, as we think about that, Craig, and this is one of the, the arguments I hear that say, yes, that's going to happen, but if AI removes the parts of your job that you have hate and you're given opportunity to do more things that are fulfilling, is that really a bad thing that we've replaced the work that you do, put it over to a machine to, to do the things that stink, for lack of a better word, and then you get to bring more of the human element to, to job role and some of those sorts of things.

So I think that's exactly right.

And that's where my brain was going.

Take those things that are a pain in the ass that you hate doing and let AI do those.

But the key is going to be don't fill up their day with more pain in the ass stuff.

Let them do something that they find more enjoyable and more fulfilling, which can be managed at the department chair level or at the associate dean level.

That doesn't require the president's office or the provost's office to get involved.

Like our administrative person in our department loves interacting with students.

Well, if we can take some of her more mundane things and help her use AI to do that, it could free up more time for her to have student interaction.

I think that's what we need to be looking at.

So there's one thing that we haven't touched on in terms of this context for discovery and that's access to tools.

So I think that's going to be a little bit of an impediment.

But I'm more and more impressed by what you can do with free tools.

So they're really pushing what you can do for free, even if the capacity is limited.

You don't get as many tokens with free tools.

But that could be a way to get this started and then if somebody demonstrates that they're willing to experiment and and gain new ways of either increasing efficiency or effectiveness, then maybe you can justify paying for the tool for them.

Yeah, I think that's exactly true.

And whether it's on the staff side or the student side, I think there is a lot that is possible with the free tools and even more when it's paid for.

But if you can look at it from a return on investment perspective, and I think about the $20 subscription fee you would pay for one of the tools, if I can find a place where I've saved someone one hour a month based on what they're doing, you functionally paid for that tool by saving one hour per month by using the $20 subscription.

So the math to get there isn't that hard as long as you're purposeful about what are those use cases we're finding where it's going to save people time so they can spend that working on something else.

I have a thought on that because my analysis amongst the user groups here at Memphis has led me to realize that the BYOD approach is what's going to work for us in the short term because of the funding issue.

But the second question I've discerned in dealing with the CIO on this is what do we do about the AI industry?

I perceive them clearly as becoming economically predatory upon us in terms of their pricing models.

Microsoft with Copilot implementation is notable.

We were able to use it across all applications up until recently when they decided no, it's going to scale back to just Outlook, which is not a bad use for it.

I'll get back to that in a moment.

But they were trying to make hay while the sun shines.

Can't blame them for that.

On the one hand because it's capitalism, but on the other hand, diffusion is going to work a whole lot better if people can find it more interesting economically supportable to do it.

I'm with Craig on the notion of we've got to provide these tools.

If we go the BYD approach, it's going to be a patchwork of people using different tools and you won't find the consistency.

You need to be able to teach people how to do things effectively is my concern.

I want to break this down because I think it's an important point.

So eventually the economics of this has to change because right now these $20 a month, even the $100 a month subscriptions, are loss leaders for the big AI companies.

There's almost no way they can be making money on a $20 subscription.

And so it's the give them a taste, get them hooked.

But I see right now the company's kind of providing more and more and more for free.

For Claude, it's almost a joke now that they're providing 2x access until some date, and then when that date gets close, the date gets pushed out again and it gets pushed out again.

Plus, yesterday's frontier models are relatively cheap compared to today's frontier models.

So we don't need to be like, I never use Fable 5 almost, because it just isn't necessary.

So I think that's one of the things we need to add to literacy, is how do you choose the right model?

But you're right, Tom.

I think in the long run, something's going to give.

I just don't know what it is.

I want to put my marketing hat on for a minute and observe.

We'll call it the Zuckerberg model.

And its monetization I perceive as being directly linked to access to the user's thoughts and usage points patterns, not the user's subscription price.

They're making money on watching us and letting us train their model.

That's not a bad thing, because fair is fair.

You get what you pay for.

And in some cases, we're paying with data.

In some cases we're paying with giving up our privacy, and in some cases we're paying with a subscription.

But you make a point.

$20 Doesn't sound economically feasible for them, even though it sounds economically punitive to us to require students to take a $20 subscription, not to use it in class.

Sometime we're going to meet in the middle with the industry and realize that access to data is part of the equation.

Just like with Amazon, for example, their ability to watch you buy is part of what makes them profitable.

They learn how to be better and more efficient simply by parsing your activities.

I think the same thing is happening with AI usage, but we're still at the baby steps phase of it.

It's almost like Google.

That's where Google figured out how to monetize the data.

Yeah, I should have the Zuckerberg model.

It's the Google model.

Yeah.

Yeah.

All right.

Any last thoughts on this before we move on to the very last segment?

Our cool uses.

Does using the word cool make me uncool?

I'm not sure about that.

There are new words for that now, and I don't have them because I know there's a generation gap and I'm on the older side of it.

Well, at least I didn't say hepcat, which is my favorite.

Go to Rob.

Will have to ask one of your kids.

We're.

So far, it's not unusual for me to get clarification on things from.

I think it's part of why you have.

You can stay up on what college students are talking about these days.

We need eye roll recognition software.

So.

Yeah.

All right.

So I've been playing around with ChatGPT's sites feature, which is really cool.

It will build a dynamic website for you.

Like Tom, you've been talking about enrollment.

I'm not saying you'd want to release your data into this, but if you did have the data available through some sort of a connector, an API, whatever, you could have ChatGPT create a dynamic enrollment management website.

And so I don't know if you had a chance to look at the.

The one I did for our AI Goes to College daily briefing.

It's impressive.

And I'll put a link to this in the show notes.

It's a publicly available site.

The design is good.

What happens is I have something called a scheduled task that runs every weekday that goes out and scans the AI news that's most relevant to higher ed.

And it creates a briefing, puts it into a markdown function file in a Google Doc that Rob and I share.

And I just said, oh, I wonder if I could just create a dynamic website with that through this ChatGPT sites.

And it did.

And it.

I don't know, it's pretty impressive, I think.

Yeah, it is pretty cool.

As someone who's looked at our briefings in traditional formats and which have been great.

Right.

It's goes out and did a lot of work and put them into a very accessible place.

This is even better.

And it feels professional and is a lot easier to digest and to read.

It's interesting.

At the same time you sent me this.

I'd been working with an undergraduate student on creating a professionalism website.

So we're working on Ask the question, what should students wear to various different engagements with recruiters?

And so there's business, formal, business casual, all these things.

And what do those words mean?

And we actually created an interactive game using this exact same tool that helps students to make choices and get feedback on what were the right choices for different activities.

And it blew my mind because it just was so easy to create something that pre the existence of this, you would have had to have been quite a web programmer to be able to figure out how to do well.

And a designer.

All of us could have.

Back in the day, we could have slapped together some HTML, but it would have looked like garbage.

This thing looks pretty good and it was straightforward to create.

So I could see something in a course where.

All right, Rob, let's say you're teaching cybersecurity where it could go out and ChatGPT and Claude can do this.

It doesn't work as well.

The scheduled task thing works pretty well, but the artifacts in Claude aren't quite as good as ChatGPT sites right now.

But you could set up this automation that would go out and look for data breaches, look for zero day exploits or whatever it might be, and build this website for your course and tell students, hey, before you get into class, you need to spend five or 10 minutes looking at today's cyber news, because that's what we're going to open class with.

And you could do it.

I spent, I don't know, maybe a half an hour creating that.

The briefing took about another half an hour to do.

But it's mostly just.

It's a chat interface, describe what you want and it goes and does it.

So I can see tremendous use that could be engaging for students for this.

Well, I could also see even in an intro course, empowering students to do this themselves.

Right.

You show them what's possible, maybe because you're using it, and then they have assignments where they go out and actually make really cool websites and, you know, with guardrails.

With guardrails.

Not NSFW stuff.

They're absolutely with guardrails.

But.

But one of the things I found when I've worked with students who show up from all different disciplines is by the time they're done spending a day playing with AI, they walked in a little bit hesitant, maybe afraid of the fact that they were going to be asked to do technical things and end the day blown away with what was possible and what they were able to accomplish in a very short amount of time.

So this is, I think, empowering of young people.

You've hit on something that I think is a big meta point.

AI lets people do stuff that they couldn't do otherwise.

I was playing around this morning in between tasks and created four pretty good songs about border collies.

I had Claude write the lyrics, plug the lyrics into Suno and, you know, gave it a prompt, wanted a bluegrass beat.

They were, look, are they ever going to be top 40?

No.

And I would never publish these things.

I'm not, I'm not a songwriter or a musician.

But before the best I could do is walk around irritating my wife by trying to make up stupid songs.

And so I think a dynamic website like this artwork, maybe we don't want to call it art, but whatever it is, AI lets us do a lot of things that we couldn't do before.

You're in my wheelhouse now.

My avocation is music.

I'm the founder of the Memphis Blues society.

Back in 2005, I think.

Allow me to interrupt.

Yeah, he's being very modest.

Well, go ahead and tap into bluetattoband.com that is a site I generated with AI because I can't build a website using the modern tools.

But Wix.com it sounds like a free commercial.

I was looking for a place where I could get a domain where I could do content management, where I could build the site.

And basically I'm running my website with AI now.

It is slow, just like Power Bi is slow using Copilot.

But I don't mind so much because I don't have to worry about cascading stuff, style sheets and all that other stuff.

And it got me my domain, which was the main thing I wanted the domain that matched the band name.

And I was just tickled pink that that all worked out.

Plus I've got it on my laptop where I can just pop it open and modify the website on the fly.

So that.

That's, I think, for a geezer approaching retirement who's not going to be an AI guru.

Me.

The ability to do things that I couldn't have done otherwise.

Power Bi build a website in the modern model made AI precious to me.

And so I'm.

I'm not the Luddite I was before.

I'm a big fan of, as Craig said, using it to do things I couldn't have done otherwise.

There will be a link in the show notes.

Well, there's the Cooper Young Sessions is the album we're getting ready to release.

And that's about 3/4 original content.

And I think it's.

It's fair to partly cloudy.

I'm not a particularly good gifted singer, but I've been playing guitar for about 50 years.

We have fun with it and then we're way off topic.

But it is an AI built site.

As a bonus for you listeners, unless Tom changes the website.

If you go to bluetattooband.com you will get to see a picture of Tom in shorts.

This is Memphis, man.

It's going to be 100 degrees today.

And then later.

Studios are not comfortable places.

They're closed, sound isolated.

The air conditioning doesn't work that well.

No, it's awesome.

It's awesome.

So, all right, before we wrap, any last thoughts?

I'm just happy to be here and I appreciate what you guys are doing because people like myself can go quite a bit faster and quite a bit further with what you're providing for us as part of the learning and diffusion process.

It's a valuable service you're providing, and I encourage everybody to check out all the episodes at Your favorite podcast, AI.

Goes to college.com and on the note.

With the theme of this podcast, Craig, one encouragement I would like to share is as faculty are entering the start of their semesters, it's amazing to think we're on the cusp of another academic year.

If you have exciting use cases that you haven't heard us talk about or you want to share with us, feel free to shoot us an email at the email addresses Craig will provide here in a minute and share with us what's going on.

We're beginning to gather some of these sorts of things and let's not reinvent each other's wheels and let's find a way we can share.

Did you not remember your AI Goes to College email address?

That's Rob Crossler.

Mine's Rob Crawford.

AI Goes to College.

I just can't remember yours, Craig.

Mine's just.

I can't type that much.

Mine's just craigai Goes to College dot com.

Perfect.

All right, Tom, thanks for joining us and thanks everybody for listening.

Thanks to another episode of AI Goes to College.

We will see you next time.

Thank you.