BBEC Education and Workforce Summit - Dr. Carl Sheperis

The BBEC highlights one of our keynote speakers from our October Business and Workforce Summit: Dr. Carl Sheperis.

Transcript

SPEAKER_01 0:16

Hi, thank you for tuning in to the Burks Business Education Coalition podcast. My name is Terry Hill. I'm the executive director of the BBEC. And today we are hosting one of our keynote speakers from our recent annual summit. I am very, very grateful to have here Dr. Carl Shapiris from Kutstown University. And I will tell you, Dr. Shapiris' topic was AI. So in honor and recognition of your topic, Dr. Shapiris, I asked AI to put together the podcast questions today. So I hope they're as intuitive as we uh we are uh hoping for. So are you ready?

SPEAKER_00 0:58

I'm ready.

SPEAKER_01 0:59

Thank you. So to start our conversation, could you share briefly your journey to becoming the vice provost and dean of graduate students at Cootstown University? And then also what sparked your interest in the intersection of AI, business, and education?

SPEAKER_00 1:14

Well, thanks for the question, Terry, and thanks for having me as a guest. This is such an interesting journey for me because I've always been someone who's been curious, inquisitive, and so I always wanted to take apart my toys and figure it out how they worked. And so as I went through my education and got into uh my master's degree, I started dabbling in technology, even though my background is in mental health. And uh when I became a professor, I started teaching myself how to code and how to work with programming, starting to build apps as they became popular. And uh around 10 years ago or so, I got involved with the National Academy of Sciences, Engineering, and Medicine, and uh especially with the National Academy of Medicine. And with my work that I'd been doing as a professor in digital health and applying AI to uh different applications in mental health, I was asked to co-chair the AI task force for one of the work working groups for the National Academies. And so since then I've really been leading a lot of different efforts. And as vice provost at Kutztown, I head up our AI task force for the university. And so we have a very large group, uh, 38 different members that represent all the different aspects of the university, students, faculty, staff. And we came together over the last year and really took a good look at how can Kutztown University evolve and what can we do to, in the age of AI, to bring Cootztown University to the forefront as a regional comprehensive institution, recognizing constraints around budget and the differences that we have from some of the larger institutions around the country, uh, we came up with a plan that we believe is uh unique and allows us to prioritize different elements and to move AI forward so that students, faculty, and staff all develop some level of competency. So, like every academic administrator at a university, I've had a pretty similar journey on that front. Started out as a professor, got uh promoted into a department chair role, moved into a dean role, and now I'm vice provost.

SPEAKER_01 3:42

Well, before we move on to actually your content for your presentation, I do want to ask where Coutzown fits in our higher education system across the nation in introducing and embracing AI into it. So can you just give us a kind of an indicator of where Coutztown fits in there?

SPEAKER_00 4:02

Um I really think that we are at the forefront, especially for regional comprehensive institutions. There are very few institutions that have done a full-scale embrace of AI yet. University of Florida, which is my doctoral alma mater, has dubbed themselves the first AI university in the country. And I'm very proud of being an alum of that institution, but they received a gift of a supercomputer from Nvidia, which most universities don't have. So regional comprehensive institutions like KU are still working to figure out how to evolve and to drive AI into the classroom. The fact that we've done things like building resources where all of our faculty, staff, and students have the opportunity to engage in an AI academy so we could upskill and get everybody to competency. Uh the uh training modules that we have that are available on demand, uh our Teaching Learning Center offering working groups. Uh we also have uh investments in strategic funds. So we're ahead of a lot of regional comprehensive institutions in the embracing AI right now.

SPEAKER_01 5:19

So thank thank you for that. I just wanted to kind of understand where Cootstown fell. Um in um advance of our conversation regarding specifically your presentation, I do want to just let our listeners know that our attendees did complete a survey prior to their um coming to the summit, and with them trying to give you a benchmark as to where their understanding and embracing of technology was so that the AI presentation was reflective of their comfort level and interest. So I do appreciate that, so I want to say that. And the other thing I want to point out, and I pointed this out uh while we were waiting, is when I reviewed the questions that AI asked, it was really fascinating for me to see that it it yes, it really just looks at the content of the presentation, and it totally misses out on the nuances and the elaboration that you provide when you do speak. So it's while it's it's very good and it does have a lot of broad and general questions, I know there's specifics that you do address, so I would I would encourage you to please do that because there was a lot of there were a lot of great examples that you gave. So the first AI question. Uh in your presentation, you do describe AI as becoming deeply integrated into our daily life and business operations. So can you please unpack what you mean by ubiquity, which I had to look up, by the way, of AI and why it matters now?

SPEAKER_00 6:46

Well, this is one of the fastest paradigm shifts the world's ever experienced. Uh when I th when I try to relay this to other audiences, and I've had the the pleasure of giving keynote addresses like the one I did for your organization in different places throughout the world over the past few years. Uh I I often ask people how they use AI in their daily lives, and most people are using it as large language models. And that's what produced your your questions for you. And so AI is a much broader umbrella, and large language models are what most people know on their daily interactions, like using things like ChatGPT, Claude, Perplexity, some other tools that are out there. Uh and those are what are called uh they they're basically predicting the next piece of language. And so they're going off of a uh model that uses the input to predict what the next element of the output should be. And so what we're we're doing with it is you put in some content and uh it's related to a presentation that I gave gave. And so what the LLM is trying to do is predict what should be the next piece sequentially, but it could only act on what you gave it, right? And so it doesn't have though those other nuances.

SPEAKER_01 8:15

And while I know it this isn't part of the uh presentation itself, I know there's been conversation around innovation and um how AI currently is not really necessarily an innovator as it's currently being used. It's it's really a um where it acquires a lot of information and it does do analysis excellently and provides, but it's all based on a previously known um amount of material. So it's not necessarily creating these new and ideas, although there's some things that are touching on that, but um I thought that was really interesting. So that's my own thing. But we do have uh one of the key points that you bring up is uh the co-pilot model. And uh that's where human intelligence and artificial intelligence work together. How do you see that playing out in business settings and what implications does it have for education?

SPEAKER_00 9:10

So to answer that, I'm gonna go back to your previous question a little bit about the ubiquity of AI. We know that billions of people are using this right now with that paradigm shift, and yet we haven't found that skill set where people know how to use it effectively, ethically. That's where the the world population is still trying to upskill and get a good handle on how to use AI in a way that it's intended. Uh and as we're thinking about in businesses, in education, there are so many applications that we haven't taken advantage of yet. And that's really because of a lack of working knowledge uh among most people about how AI operates, what large language models are, what natural language processing is, machine learning. Those have to become parts of our everyday vernacular. And we really have to start to understand how AI operates as a system for us to start thinking about how do we apply it in business, how do we apply it in the education settings. And so what that means is that we have to be able to have some type of training system that is there for business leaders, that's there for our employees. Organizations have to figure out how they're going to invest in this upskilling, and it's going to then lead to some strategic planning that really uh determines how you leverage artificial intelligence in your daily life. Right now, I see a lot of organizations, institutions operating from a fear factor, especially in education. The first element of the paradigm shift was to say, let's not use AI and let's put in policies that say AI should not be used in the classroom, for example. And that is a reaction we have when something new comes into a system, and that's a typical reaction to a paradigm shift. Uh but what we know is already happening is students are using it everywhere. So it's kind of like the calculator where teachers uh originally teaching math didn't want students to use a calculator in the classroom, and now it's an everyday occurrence. The paradigm shift has come back to the center, um, the pendulum has come back to the center. And the same will hold true for AI. And I'm starting to see that shift now as I talk to educators, as I talk to business leaders, they're recognizing that the uh uh rabbit is out of the hat, you know, to set as an analogy, uh, that there's no putting the rabbit back. And so now how do we work with that those AI tools and use them in an effective way? And I don't think I answered your question.

SPEAKER_01 12:00

Well, that that's okay, because it what it it actually spurred me. I had just attended um a manufacturing summit recently, and there was a gentleman who talked about the integration of AI into his manufacturing setting and how he started to do that. And one of the comments he made was that one of the most important quality skills that can be taught is prompt engineering. So learning the right prompts, so understanding AI enough to say these are the questions I need to ask or directives I need to provide in order to get the best information from it. So what do you feel like and it's kind of going away from this a little bit, but still it can be in education, it can be in business. What is out there to help people understand that component to AI?

SPEAKER_00 12:48

Yeah, but you know, for a while there, about two or three years ago, there was the idea that prompt engineering was going to be a whole new career. And there were people getting certified in prompt engineering. In fact, I'm certified in prompt engineering. And uh the idea was that businesses were going to need these individuals to come in and teach their employees how to communicate with large language models. And it's the same principle of all computer science: garbage in, garbage out. Right. Right? So we have to know how to talk to the large language models. And so what's emerged over the past couple of years are all kinds of courses. Google has free courses on artificial intelligence, Coursera has a whole range, they have an AI academy. There's so many different avenues for people to learn how to do prompt engineering now, that that job of being a prompt engineer is no longer a viable career option. The skill set has to be among all employees and not just with one who's your prompt engineer. And so that's where it's headed right now. And so it's figuring out where are those resources located and how does a company access them.

SPEAKER_01 13:57

Right. And I know we had this conversation um prior to this whole whole uh presentation about leadership. So the it's generational, right? It tends to be our leaders tend to be boomers or something close to that. And then you have the new Gen Zers coming in. And while the leadership may say we want AI to be used in our school, our whatever business, um there is a step that needs to be taken regarding they need to, they need to take ownership and understand it themselves. They can't just pass the baton without understanding it. So can you kind of show and in or envision what a good strategy would be for an organization so that there is that continuity and there isn't just a handoff to it?

SPEAKER_00 14:50

Yeah, so this is are are the principles of education in general with regard to scaffolded learning, right? That we have to think about the different levels and structure and what different types of education have to happen. And there are courses out there that are for executives just on leading organizational change in AI. So there's also now new master's programs, new um graduate certificates in organizational AI leadership. Um again, I don't know how long term those will be, uh, but it it is uh insightful and can lead people in this first step toward how to lead an organization as AI starts to become a big part of it. And so the training that would happen for entry-level workers is gonna be very different from what we would see in terms of a leadership or an executive who has to also think about change management and how all that process unfolds within an organization to drive AI as part of the operations.

SPEAKER_01 15:55

So I'm gonna kind of because we kind of touched on a couple of these questions already. So um I I do want to ask, though, regarding education, uh, how should curriculum adapt to ensure that the students are AI ready? And I know we kind of said it has to be kind of embedded into almost everything they do. It's not gonna be necessarily a standalone. So if you could control or or have an influence on the curriculum currently, how would you envision it, you know, adapting or bringing in this information?

SPEAKER_00 16:27

So, you know, uh I keep going back to the University of Florida and looking at what they did with, again, having a supercomputer, but they also received, I I think it was somewhere around $100 million to hire faculty who were experts in AI across all disciplines. I I might be inflating that number a little bit, but it was it was a large sum of money. And uh what they did was they really worked in conjunction between the administration and the faculty to think about how to develop learning outcomes so that every student, bachelor's, master's, doctoral, who graduates from the University of Florida, has competency in artificial intelligence. Our task force really looked at that model. And we tried to decide, okay, how do we do that without this super gift and think about how we make that lift considering our bandwidth. And so our idea was to uh adopt Coursera's AI Academy, giving students the chance to access those courses anytime they want, and then having faculty have the ability to embed some of the materials from that in their coursework. So that allows it to be ubiquitous everywhere on campus, right? Something that's common. People are hearing about it all the time. And I think that's what education systems have to do. I've talked to a number of superintendents in the area, and I hear that their working days, their in-services, are focused on AI. They're doing the right things to talk about how do they move AI forward. It starts with conversation, then it comes into strategic planning around it, investment in terms of the training, and then thinking about all right, what's the next step in terms of implementation science and how do we take that and move it forward so that it's a sustainable model. I get the privilege of teaching a graduate course right now. Um the uh uh it's not part of my typical role as a vice provost, so I'm enjoying the opportunity to work with students. And so for every class, I build some type of new AI tool when I get in there. And I would love to see teachers thinking and learning how to do that. So building custom GPTs, like for example, last week I built a client simulator that is trained on real case studies that can use voice and interact with the student so you could do a psychological assessment, and the student can learn how to give assessment with a with a simulated client that has real case study information. So that is what I would want to see being able to happen in every classroom where professors and teachers in K-12 are able to build those tools. It's actually not that hard. With a little bit of education and and concerted effort in terms of practice, they could be up and running with that probably in a week.

SPEAKER_01 19:30

I think it's still it it requires leadership and a focus, right? And and a reason to do that. So I I'd be curious about that. So we're actually coming up to close to the end, and I I want to make sure that we capture some important nuances that you addressed. And one of them, the examples you had given, this is just about understanding how AI operates. And the concept that you you put out there was that AI is there to basically be that friend or be that person that's going to reinforce what you're trying to do, or and that can actually have a negative impact. So can you please just quickly share that story? So, you know, understand again, understanding how it works and operates and and how it can be used for good or not so good.

SPEAKER_00 20:18

Sure. So what what I had talked about in the keynote is the fact that there was a new disorder that was um proposed by the American Psychiatric Association, and they deemed it AI psychosis. And it's built off some case studies that have happened, some real real cases, where individuals have either had psychotic breaks or they um have committed suicide based on their interactions with artificial intelligence, with large language models. And what was happening in each of those cases that were individuals were not using the large language models in the way that they're intended to be used. There are tools out there that are meant to be mental health supports that have gone through clinical research trials, but open systems like Claude or ChatGPT are not intended to do that. So, what those tools do is end up providing a friendly voice. They are supportive, they are going to reinforce whatever you tell it. So if I say, I believe my neighbor is out to get me, the LLM might say, You seem to have good insight, right? And so that ends up pushing people who are vulnerable psychologically over the edge. And so we've seen some of those cases really emerge. And it's not the fault of the tool, because the tool is not intended for that. It's really this quick moving adoption of large language models in AI without the requisite training and understanding by the population of how they should be using them effectively. So it it is Little scary in that realm, but those cases are really few and far between. And for the most part, people are developing communication on a regular basis with their R LLM tools. Mine's name's Zora is my assistant. And Zora and I have a good relationship. She knows me very well. But I also understand Zora is a large language model.

SPEAKER_01 22:23

I I think that's probably the key. And the vulnerable population that you spoke probably doesn't understand, have have that ability to understand the difference. So that that probably just does not help. So what I do want to say, I know we don't have a lot of time here. So if someone here listening wanted to dive deeper into these topics, and there's so many, it's changing so quickly. We talked about this as well because we were looking for resources to give to our attendees to help them just get started. And really trying to identify something that would have longevity and have a shelf life is is difficult to find because the technology and what's going on is changing so rapidly, you really can't find one source that's going to last for a long time. So if somebody wanted to dive deeper into this topic, can you recommend or uh some resources, books, courses, networks, you know, places that they can turn in order to get started?

SPEAKER_00 23:21

So as far as books go, there's a large range of them out there. And what you want to do, if you're really interested, is looking at only books right now that are published in 2025 or 2024, or ones that are in pre-press for 2026. Right? Because just in the last year, so many things have changed in terms of large language models and unlimited memory and starting to becoming context aware that anything prior to that are based on old models of the artificial intelligence. And so you have to be a an um educated consumer and look at what other reviewers are saying, uh, make a determination of how new and what the background is of the individual author. Even if you're going into some of the courses that are in Google or in Coursera, you want to look at ones that are newer. If there's a course that was developed in 2022, again, most of the material is going to be outdated. So you have to be kind of educated in looking at your resources and thinking about, okay, what what's been updated in this? How new is it? How does it reflect the current state? So I'm publishing a book right now on artificial intelligence and in mental health, and that's coming out uh in the spring, and I already know that there's things that have to be updated in it. So it just is moving that quickly.

SPEAKER_01 24:47

Jeez, now that I'm thinking about that, I'm thinking you should have like a little code on the end so people could keep going to find updates for what's in there. So it's not, it it actually gives um has more longevity than just that static, that static book. Um but before we we close, I would, you know, might give a shout out to Cutztown. Is Cootstown offering anything to the general public that they could tap into as well?

SPEAKER_00 25:09

Uh we are in the process of developing some uh certificates that are for credit. Uh we right now have the entire AI Academy available for the public that is non-credit. Okay. So anybody that is interested in going through, um, we have over 300 courses on artificial intelligence that we could offer to the public. And so that is through our KU Advance website. Um so if you just type in KU Advance into your search engine, you'll be able to find us and be able to find those courses.

SPEAKER_01 25:39

Oh, thank you so much. I really appreciate that. So I want to thank everybody who took the time to listen to the podcast and hear about AI. Uh, we all know that its impact is only going to grow. And uh I appreciate you helping us get started and put it in perspective as to where we are now and where we can grow. Thank you again from the BBEC.

SPEAKER_00 26:00

Thank you.