Hi, good afternoon. Thank you for having us today. We a few months ago did this panel. We hosted the Midway Chamber of Commerce at St. Paul College and we did this panel and that was somehow how we uh we earned an invite to come to your Chamber of Commerce meeting today. So, thank you for having us. So AI is changing the way we learn, work, and connect. And so today we're going to hear from higher education about how AI is showing up in classrooms, in the workplace, and in the community, and what it could mean for your organization. So something that we we part of our role at St. Paul College is preparing students for uh the workforce. So as you're looking to hire hopefully looking to hire our graduates, if you don't know anything about us, touch base with me afterwards and we'll chat. Um, as you're looking at hire our graduates, we understand that that AI skills, use of AI, understanding how it can, should, might be used is something that is emerging in the workforce. And so, we need to make sure that not only are we upskilling ourselves as professionals, as faculty, as leaders at St. Paul College, but how are we adjusting, modifying our curriculum to ensure that we are training students to be prepared for the influence that AI has on their individual uh career area and industry. So with that, I'm going to do a quick poll by hand. And so I'm going to ask a question and your options are not at all, sort of, and yes all the time. So a question is to what extent is your organization currently using generative AI tools? So if you say not at all, raise your hand. Okay. What about somewhat? We're we're dabbling. We're experimenting with it. Okay. And then what about a lot? I'm using AI every day for for our work. Okay. So then the second question is more on a on a personal level. To what extent are you currently using generative AI tools? >> Uh not at all. >> Are you asking how we're using it? >> Yeah. Like or sorry. Yeah. To to to what extent you are currently using generative IIA tools like personally. So not at all. Is anyone like I don't even know how to use chat GPT. Okay. What about somewhat? Mhm. And what about extensively? I use it all the time. All right. All right. Thanks. That just gives us a sense of where folks are at in the room. So, with that, I'm going to turn to our panelists and I'm going to ask each of you to introduce yourself and just touch on what your relationship is to AI. How does it show up in your work and how do you use it or not and why? >> Hi everybody. I'm Mindy. I've been teaching at St. Paul College for 18 years. So AI is a shift for me thinking about how I teach and what it means for my students. But as a business department, so I teach in the business department. Uh we have programs in accounting and marketing and human resources and also a transfer pathway. We have really had to spend a lot of time rethinking our curriculum and what are the skills that students are going to need when they graduate. Um so I think that's how it shows up for the most part is thinking about in our coursework what we need to do to train students to use AI. but also to think about it ethically um and what are the other skills that are really going to be good to complement the shifts we're seeing in technology um on a personal level I do use it also so sometimes I'll design for example a discussion board question and then I'll go to AI and say um how can I change this question so that AI can't answer it and my students really need to be the ones answering it um yeah so very briefly that's kind of what I'm doing >> thank you My name is Sonia and um I have been at St. Paul College for two years and I was at Anoka Ramsey and Anoka Tech for five and a half years before that. Uh and before that I was at a K12 institution or a district um for about 10 years and before that in higher ed for about 10 years. So I've been in in academia for quite some time. Um I have taught and I have managed technology in all different uh forms. Uh currently at St. Paul College. I manage the IT group and so I see oversee both the service desk, the data center, data services, IR and the network team. So I get to see all of it. Um and when I was teaching, you know, I always saw um while I was teaching technology, I always looked at uh literacy skills as part of my role. And so as I was using AI, it was like how do I use AI as as as a literacy tool? But then using AI now it's it's just it's kind of embedded right um when we look at um web searches right now if you go to Google and do a Google search Google right um you get the the Google answer and it's an AI answer so so I use it all the time uh whether I um whether it's intentional or not because it's just it's just one of the tools that we have u we also we also use AI because it's whether it's in a software package or now even in a hardware with the software package embedded it's so it's it's in everything that we do um when students are coming to the service desk especially they're asking us how do we use AI or how do we not use AI so that's just something that we're always seeing um but I also see that AI is our is the next generation of a tool that we just have to get used to using but and I will also say too I'm a huge Marvel fan and what Uncle Ben said was with great power comes great responsibility. And so the next piece that we have to look at is how do we use it in an ethical manner. All right. I don't know how to follow all that up, but hi everybody. I'm Claire. Um and I too have been in higher ed for more than 20 years. And I will say higher red keeps you young. Um so I um have been working with AI tools since November of 2022 when chat GPT first hit the scene. Um, I've worked with, um, developing resources for faculty. That's part of my role, helping folks figure out how to navigate this world. And I am, you know, kind of a skeptic, I will be honest, but I know a lot about this tool. So, I would say that my use of the of um, AI tools kind of fall into three buckets right now. So, the first bucket is I tend to use it for things that are less meaningful to me and the organization, but that I already know how to do. So the reason I do that is because um if I already know how to do something then I have already a good skill set to say what AI is generating is good or not I need to make these tweaks or not. If I'm using it to do something I don't know how to do how do I know that it's actually generating something that's good or useful. Um so one example there is um the other day I wanted to look closely at a contract and I was looking for something that came up over and over in the contract. I know how to do that. I can read a contract. Do I want to read a 200page contract? No. So, that is something I could put into our internal tool copilot. Um, and ask some questions about what's in that contract. I could do that and I can tell if it's right or wrong based on my knowledge. The second bucket is I use AI tools when I am not worried about skill loss. So, one of the things we're learning with AI is that if we use it over and over and over again, it might weaken some of the skills that we're using AI to substitute. I do a lot of writing in my work. I love writing. That work is meaningful to me. So, I tend not to use it so much for writing tasks because I'm worried about it weakening my writing skills. So, I really think about the kinds of things that I can use it for that won't weaken a skill. Um, so one example of that is we've been experimenting with these tools to translate handwritten text into um, OCR character recognition text um, for accessibility reasons. And I'm not worried about losing my ability to read handwriting. That's a skill I probably will have my whole life. So that's one example of what that looks like. And the third bucket is um, to develop my literacy. So I spend a lot of time poking around with these tools to see how they work or how they don't. I do a lot of something I think of as equity stress testing to see what's how's it working now? Has it improved? Has it not? So that's kind of the third way that I tend to experiment with these tools. So I'll leave it at that. >> Thank you. So what skills should colleges be teaching today so students are prepared to use AI responsibly when they're working for folks here in the workplace? Oh no, you got to hear from me again. Sorry. So I think of kind of four things that um I would like students to be thinking about that we should be teaching. And I think this might resonate with some of your work too, but you can tell me if not later. So um one is just kind of the basics of what these tools are and how they work. Really understanding foundationally that these tools are predictive is essential to figuring out is this a good fit for what I want to do or not. Um, the second, and this speaks a little bit to kind of my own skepticism, is what their limitations are. Where does bias show up? What are some of the challenges with using these tools and what do they do imperfectly? So, those are kind of the standard skills I think that we're hearing across higher ed that folks really need to know and have in their back pocket when they're showing up in the workplace. Um, the third is when to use these tools and when not to. And again, I keep thinking of the skills conversation. So using tools when you know how to do something and you can make it more efficient versus using them for something you don't know how to do and when you're learning for the first time whether you're in college whether life you're a lifelong learner using the tools as a substitute for developing those skills might not serve you in the long run and then the fourth piece that I think about often is the impact of AI usage especially on your relationships. So we heard a little bit before about relationships, right? And how important relationships are in the work that we do. And so thinking about your clients, your customers, thinking about our students, what does it mean to think about using AI in relationship with folks? How would folks feel if they learn that you use AI for these tools? Maybe they for for these tasks, maybe they'd be fine with it. Maybe it would bother them. You have to know a little bit about those folks um that you serve and what your relationship with them is to decide what's the impact of this tool going to be. How is it going to affect that relationship? Does that make sense? Those of you who are head nodders, I so appreciate I am a headnodder. So that helps me. That impact piece I think is a really crucial skill because you don't want to end up on the news in a bad way. You know, you don't want to end up in a lawsuit. So those relationships pieces I think are really important to think about for our students and for all of us really. Thank you. I always I always like to look at the history of things first. Um before I start looking at the the future of things. So before looking at and starting to use AI, I always think about how did we do things before. So, kind of speaking to you as well or to some of your points, do you remember uh card cataloges in the library? Remember the wonderful musty smell, the long long cataloges and getting um paper cuts from the note cards. Um I always think about how long did it take us to find a certain book when we looked at the card cataloges or when we had to learn uh long division, right? We that was a certain skill that we had to learn. And then then came you know web searching and then came the calculators and then came chat GPT. I mean there was a few more things in there right we have to know how to do those things before we start to use these new tools. And so I think that's the other piece too that students need to understand like how did we get to this new tool and if we don't understand how we got there because we we don't understand the history it's really hard to start using these new innovation new innovative things. Um when I first started using computers I um I got to use the old Xerox machines that had like the centipede roller ball. Did anybody remember that? Okay. And the like the 13inch floppies. Yes. I'm that old. Um, and Word Perfect 3.1 with the reveal codes. Yeah. Yeah, I'm getting some. Yeah. And so, and so now when we look at Office 365 and it used to be, you know, 2021 19, etc. Those were the old tools. And so, now we start looking at these new tools and new ways of use using these tools and then we start to embed these AI tools. We can appreciate AI. And I think that's why the students need to understand before we start using these new tools, it's really important to understand like what was what were the foundational pieces in order to start to appreciate these new things. And so as I look at my team because they are at sometimes their students as well when I start to look at my team before we start to use these tools, they have to understand how do we get there first before we can start using say Cisco tools. We have to understand the code before we start using you know um Python or whatever. And I'm not a coder at all. But before we start looking at this, we have to understand DOSs. We have to understand all the all the shell codes or whatnot. We have to understand these things before we can start looking at AI tools because we have to be able to validate, right? I had somebody just recently, someone on my team recently use a co-pilot to do some um to to want to update something on our domain controller. And it looked beautiful and and he presented it to me. He was like, "Yeah, we should do this like right away." And then I had somebody else on the team who said to me, "This is absolutely wrong." And if I didn't know any better and if the team didn't know any better, we would have really messed up our entire infrastructure and no one would have been able to use any of the computers because we would have hosed our our our entire infrastructure. So AI is great, but if we don't know the history and if we don't know how to do things by scratch, AI is not a great tool. And so we have to be able to understand foundation before we can start using the the advancements of these tools. So that's the one thing I would say that students really need to understand before they can start using AI. >> I'm thinking of like repeating a lot of what you've said, but like another way that we talk about it in terms of curriculum is this idea of cognitive offloading, right? So, I have older children and they drive and I won't name her. She went to school in Woodbury. You might know her. She can't drive around without her phone up with the map happening, right? So, like she will type in, I'm going from Woodbury High School to the library. I'm like, okay, that's about three blocks. You can handle that. But the point is that there are a lot of things that AI is so easy to use and this is a challenge that we're seeing as teachers um that it's hard to get students to understand that in a real way because it is so easy. But what we're seeing as people come through is things like critical thinking and creativity are going to be really important. Right? So I look a lot at what are job projections and what are the jobs that are going to exist and the jobs um that are going away tend to be the ones that have a lot of automation room for automation. The jobs that are going to persist in areas like construction and education. These are things that require creativity and cred creative uh critical thinking and those types of skills. So, we're working really hard to get students to practice those things and relationship building at a time when they're even more kind of isolated and into their devices. Um, so it's a really big challenge and something I'm thinking about all the time. If you were advising a small business or nonprofit leader again like some of our folks in the room and you were just beginning to think about a they were just beginning to think about AI, what's the first step you'd recommend? >> Can I? >> Yeah, please. >> The first Well, so my background is also in human resources. I would say the first step I would recommend is having an AI use policy. So a policy for your employees that outlines how they can and cannot use AI. Um, so as we're doing our show of hands, I think there's a lot of organizations right now that are still trying to figure out how to navigate AI and what are the appropriate and productive ways to use it. But we know that people are figuring it out and using it a lot. And so for many employees, they're using AI to take shortcuts and get things done more quickly. Um, and we know that there are a lot of risks and challenges that go along with uploading data uh around privacy, especially that protected information within your organization. So, being clear with employees about what they can and cannot do with AI is really critical, I think, for protecting your organization. I saw a show of hands. Um, personally, a lot of you are using AI. I think that's what I saw. I wear glasses. So, um, when you do a web search, how many of you search on the web? Just show of hands. Okay, that's AI. Just letting you know. So, you're using it. Okay. Um, so I I would also say while there I absolutely agree, you need in a business to have a policy, a use policy, and protect your data. By all means, protect your data. Don't upload things. um don't upload that anything that you shouldn't be uploading, right? But it is also if you use it the right way, it should not be a scary thing. I would also say too, there's a lot of there's a lot of product out there, a lot of applications that have AI already embedded. Those applications have a specific those those AI tools are built specifically for those applications. those are trained in those the AI is trained within those applications. So you can't necessarily uplift that AI bot or whatever it is and move it into a different tool because you will get different results that you're not expecting. Okay? So just just know that when you're looking at um a tool that you know you you saw you know XYZ company was using and you want to put it in ABC company, most likely it's not going to work in the way that you anticipate. So you do have to check for hallucinations or inaccurate inaccurate um uh results. So that would be the one thing I would say when you're when you're thinking about uh looking at AI for your your business um or even personal use. >> All right. So one of the things I find really important in my life in the work that I do is to bring a researchmindedness to when I'm learning about something new. And what I mean by that in this context is you could do a little bit of research to find out what your peers are doing, what your competitors are doing with AI specifically, whether it's through conversation, whether it's through social media. Get a sense of what folks are up to. Would it work for you or not? And why? So, those are the kinds of questions to ask as you're looking around seeing what folks are up to, getting the lay of the land. Maybe you want to experiment with something that someone else is trying, maybe you don't. maybe that wouldn't work for you or that doesn't align with your values. And so as you're doing that, I think that could be a good foundation to a policy, developing a policy. Here's what I've seen, you know, someone else do and it seemed like that could be a really good fit for us. Let's try it out. Or here's what I've seen someone else do and yikes, no thanks. So, and maybe that goes into our policy. So, that kind of researchmindedness, what are folks doing? Would that work for us and why or why not? I think those are good questions to kind of keep asking yourself as you're looking around and seeing what other folks are up to. So, as a part of our time here today, we were asked to possibly give a little demo or some guidance on how you develop an AI prompt. And so, we're going to hand hand it over to our faculty member on the panel here, Mindy Travers. >> Yeah. So, I mentioned this that, you know, we're trying to get students to focus on their critical thinking and their creativity, but we also know that going into the workplace, it's important for them to know how to use AI. So, what I'm giving you is just a short version of a lesson um from one of my classes where we talk to students about appropriate use and how to use it. Um, so the first thing I would say is you can just kind of bring up a tool like chat GPT or Gemini and use it very similar to like a Google search, but ideally you're giving it what I'm going to call some context. Some people might think about this in terms of the training the tool. But we're going to think about, you know, what's the information that you want the tool to be consider to consider when helping you. Um, so if we go to the next slide, what I mean by that is that um, AI tools, generative AI tools, they're not smart, but they're really, really fast, right? So if you just plug something in, it's going to give you something back. So the more that you can kind of direct it to give you meaningful um, information, the better. So in this example, um, we're thinking about it in in the context of working in human resources. So hopefully you would already have some level of expertise in HR if you were going to use it in this way. But you might start by telling the tool what to do. So here the exact exact example is you act as an AI assistant for an HR manager of a private business. Your role is to provide clear, accurate, and practical guidance that is deeply rooted in US employment law. So you're telling it this is the kind of information that I want from you. Um, and you can also let it know what sort of tone you want it to have. So here we're saying maintain a professional, empathetic, and expert tone at all times. Empowering users with reliable guidance and actionable advice. So before I get started in asking any questions, I might give chat GPT this direction. Um, and then for the next Yeah. So this slide then we also, you know, it can tap into any information, right? these large language models are gigantic, but maybe I want to know that I'm getting reliable information. So again, here I'm saying please use the um Equal Employment Opportunity Commission website as one of your primary sources. So I know as an HR person when we're thinking about law that that EEOC site, for example, would have really reliable information and that is where I want the primary source to be. So I might include that, right? and say when a query can be answered using these materials, reference concepts, examples. So give me that and then also include the hyperlink. So I might also again direct that chat GPT tool to say when referencing external resources include the clickable hyperlink when appropriate and then that means that when I get an answer I can verify. So I can look at it and go does that seem right? It seems right. Is it really right? So that you get that sense of exactly where things are coming from. So we would give that contents be t context before we'd even ask it anything. Um yeah. All right. If we go to the next slide. So then in terms of prompting uh I like to talk to my students about different types of prompting. So one is called tree of thought. And you might be working with a tool that already knows what tree of thought prompting is. Um but the idea with tree of thought is you're really asking that tool to give you multiple possibilities. So one thing that a a tool like chat GPT likes to do is give you an answer and and say this is the one answer and rarely is that the case. So here you're kind of saying to it can you give me three different options so that I could understand like what would be three different approaches and what are the benefits and challenges that go with each of it. Um, so for example, if we were thinking about I have an employee that's struggling and I want to do some coaching or I'm going to need to write some sort of performance improvement plan. Could you give me three different approaches to how I might handle this? And then again, I can draw from my own expertise as a marketing person or an HR person to say which one of those really does sound like it's the best. Um so in that type of prompting that's really what we're doing is saying give me these three different branches or a number of different branches and identify what is the thought process that goes with each of it so that when you're pro uh provided with a solution from the tool you have that ability to tap into your own expertise and say which one really aligns with our inst our company's values or kind of um my own expertise and what I think is correct. So that tree of thought prompting is one way to make sure you're not just kind of getting one answer, it looks good and going with it. Um, and then so this is I've given an example exactly what that would look like. So what I would type in is not just like what do I do with performance improvement plans? You would say please create an improvement plan using the tree of thought approach and maybe I have had to tell it exactly what that is and the main objective is to improve employee performance. So, I'm not saying I want to get this person fired. We're, you know, I want the tool to know what I'm trying to do with it and what are three approaches that could be used to write a performance improvement plan. Explore each solution, including possible advantages and disadvantages of each approach. So then when I'm getting this, you know, I might get a performance improvement plan that looks really good and something that would have taken me maybe eight hours to write, it might just show up instantaneously, right? So, it can be helpful. I would not advise you to do this kind of thing if you have no expertise in HR and you've never seen a performance improvement plan before, right? Um so again, this holds true if you're thinking about marketing issues or um accounting or whatever it might be. The next slide talks about a different approach which is called chain of thought prompting. And this is where instead of looking at multiple options, we're going to go really deep into kind of one idea. Um, so again, it's a technique that you can use where you're kind of giving a logical sequence of steps and telling the tool exactly what you want it to do. So this involves that idea of breaking down kind of a more comp complex task into these um smaller manageable parts. And what it allows you to see is like how is the tool thinking about this answer? And again, we come back to the idea of needing to assess it. Um, is it a good answer? Is it something that I should go with? So if you go to the next one, I've given an example of kind of what this might look like. So here again, we're still thinking about creating a performance improvement plan for somebody, but we've laid out nine very specific steps. So I'm saying to the tool, define employee improvement plans. So I know that it we have the same idea of what this is that I'm asking for. Um explain why an employer would use a performance improvement plan. Identify the main reasons an employee might be given a per performance improvement plan. Provide some examples of a PIP and then describe the major laws that need to be considered when I'm developing this PIP. Discuss the ways a PIP can be discussed with an employee. Explore potential benefits of each approach. Provide a script that can be followed when delivering a PIP to an employee. And summarize the key points discussed in this chat. So then when I'm getting again all of that back from it, I can kind of see what are all the things that I'm going to need to know about to evaluate whether or not this performance improvement plan that I get at the end of it is worthwhile. And there are some things in HR like writing job descriptions and you mentioned contracts, you know, those things that take a lot of time um where this can be really helpful and more efficient, but you don't want to get this stuff wrong either. You'll be in a lot of trouble. So that's one example of kind of what we're talking about in some of our courses and I thought it might be helpful since we got that request. So thank you. So with that we wanted to open it up for questions. >> So one question we get question is how accurate is the response when you're doing research just the general searches? How accurate is that? I realize it depends on where it's coming from and what you're researching. Do you have any statistics to tell us more about that? >> Boy, I wish I could say yes. I can tell you it is 63.7% accurate. I think the reason why that I so appreciate first of all full disclosure you asked this question before the panel so I've had some time to think about it we were chatting about it a little bit in advance and I I'm going to say the most annoying um academic thing ever which is it depends but the way that it depends kind of matters. So it matters the tool you're using, what kind of information is in the training data for the tool, and whether the question you're asking it to address is likely to be represented in that training data. So if you are asking it a question, so when I think about kind of in a higher ed context, most kind of um tools like Chad GPT probably have a pretty good understanding of like a biology 101 course and what's in a course. When you get to more advanced courses, the literacy of those tools might go down quite a bit. And so thinking about whether what you're asking is likely to be well documented in the data the tool has access to is really critical for answering that question. If you're answering really asking really specialized questions that are very particular to your context, unless you've given these tools a lot of information about your context, which can up the accuracy rate depending on the guard rails you set up, um it might not be very accurate at all. So rather than a percent, what I would say is it depends on the kind of question you're asking and whether what you're asking is likely to show up in the tool itself. Does that make sense? Okay. Again, the head nods, bless you. Yeah, >> there is an NPR um when when uh chat GPT first came out uh hit the news and all that, there was something that NPR had about uh three hospitals that had that were training um language models and uh when they were trained on these specific hospitals, the the end results were good, right? And then they moved these language models to the different hospitals and then the results were bad, you know, relatively bad. So again, it it really depends on how they're being trained. I would also say too based upon what model they were using. For example, um I wanted to um take something that was being translated from English to Vietnamese and English to French because I've got family in Vietnam and France according to um u book language, right? The translation was perfect. But then when I gave this Vietnamese and French translation to a native Vietnamese and French speaker, they're like, "Oh, these are probably not the right way to say it." Right? So again, it was it was grammatically correct, but not necessarily how a native Vietnamese or French speaker would would say it. So again, you get some good stuff, but you know, it kind of depends on how people are going to use the words. >> Yeah. And students struggle with this a lot because very often I can tell when they've used AI and it is inaccurate, right? But that's where I'd go back to kind of that idea of how you're prompting and how you're using it too. So if if students have known to direct to use EOCC website as their source, they're probably getting reliable information. If it's more just like a general search tool, they're using it. We don't really know where they're drawing the information from. >> Yeah. And I think um brown brown shirt first. married. >> Yes. Give it to the retiring the man of the hour. >> Uh being so infant as it is, is it getting smarter? >> So is that becoming >> according to the movies? Yes. >> I was going to say according to movies. Yes. Iroot. Uh what else was it? Iroot. Skynet her Megan Terminator. >> Oh, Terminator. Right. Right. Which which one? Which one? >> I mean, I think it is as it becomes more mature. Right. We've started out with chat GPT 2.0. Now, I think we're up to 5.0. Um, with St. Paul College and Minstate, we're using co-pilot. And so, we're, you know, we're training it based upon what we're putting in there. So and and we can train it because like my example before where where one of my staff members was putting in information that was act you know and the end results were bad we were able to say no you were actually incorrect and then you know co-pilot was like oh you're right we are wrong you know we or we yeah so they can be trained and it is getting a little bit smarter but again I don't it's it's not going to take the place of humans >> it again I just have the most annoying things to say I apolog apologize in advance. It depends on what you mean by smarter. So, you know, we use the term artificial intelligence. What does that really mean? Does it mean we it does things faster than us? Is that what being intelligent means? Does it have access to more information? Is that what being smart means? And so, it's worth thinking a little bit, I think, reflecting a bit about again what functions we want these tools to serve. Um, one thing I'll say in terms of large language models like chat GPT and other chat tools, they're kind of bumping up against the limit of their programming right now. And until there is some major innovation in the way those tools are structure structured, what I'm hearing in the research and what I'm seeing kind of on the ground is there's a little bit of an upper limit to what they can do right now. Maybe that'll change. Maybe we'll get to artificial general intelligence. I don't know when. We're not there yet. So, I would say if you were like, "Oh no, the world is ending." Don't panic. We're not there yet. But if you want to talk about the um the social media site that's been created for AI agents, we can chat about that later. That's kind of an interesting new innovation. Um but yeah, really depends on what we mean by smart. That's again one of those annoying things. Yeah. Yeah. >> I also want to remind everybody that this was built by humans. >> So don't forget about that. So this was built by human people. Humans will always be smarter. I believe >> that was a good pause. >> So, I got a couple two few quick questions. Um, when it comes to the prompting, um, can you like if you have a subscription to chat GPT, can you make those those prompts on how you want things to be done once and then every time you go in it will remember that or do you have to do it every time you're doing a search? And my second question is, how do we keep the kids literate these days? So, how do you how do you at at the college like you talk about, you know, catching them if they're cheating, if they're using it, you know, that kind of stuff? Because I'm seeing this. It was already bad enough with the phones and with all the stuff that was going on before this. Now, it's it's I mean, it's getting crazy. So, how do you keep the kids engaged with staying with the skills that they really need? >> It's a good question. It's a good question. >> Well, first to your first question about the prompting. Yes. So, you know, I could set this up and say, I always want you to serve as like this little a HR assistant to me and these are the sources I want you to use. And then it remembers that. So, when I'm coming back to it, you know, we interact in the same way. Yep. Um, and I don't have an answer to the second question. This is something we're just thinking about all the time. But I will say there are a lot of students that they're kind of resisting AI in a sense. Like they want more face-toface interaction. They do want to be learning. I think what is hard is right now it's really demanding to be a student and a lot of our students have families and full-time jobs and if you have two hours to study and you really need 17, it's really hard not to take the shortcut that is a AI. Um, but we're always looking for ways to say how do we rethink the way that people learn? Um, it might mean going back to some, you know, practices and assignments that I thought we'd never go back to. Moving away from things like papers and saying you need to stand up and give an impromptu speech because I just want to hear from you and do you is this you know your thought process and um so we're shifting I think all those practices that we learn but we have not solved it and it continues I think to be a real issue especially as we've seen a spike in the demand in like online learning. I think that is the area where it gets the most challenging. >> What about the kids in high school now? Because you guys are with adults, right? Most of your teaching adults. >> Yep. A lot of PS. Yeah. High school students do students. >> I can see where adults might be. You know, they're got kids, got family, got all that, but have a little bit different approaches. >> And you have your digital natives, right, who grew up with technology, so they might embrace it, too. But yeah, I I think the best approach with students right now is to show curiosity and have a conversation. I think what we know um doesn't work super well, at least this is my observation and other folks can disagree. Um is to show up in a classroom and say you can't use AI, good luck. Um to show some curiosity for if students are resistant or not, if they're excited, why? What's their experience? Um what do we know about these tools right now? How do you want to use these tools in this class? maybe we'll experiment with it in a couple of ways in a structured and kind of um educational context. So to me I think and maybe this is just the sociologist in me speaking to engage in conversation and to approach these conversations with curiosity rather than judgment which I think is difficult for many of us because I don't know if you all feel the same. I'm tired. I'm tired and I'm grumpy. So how to approach conversations with students with our kids from a lens of wow that's really interesting tell me more about that um or oh where where did you hear about kind of that you know limitation of AI let's talk about that or let's decide collectively how we want to engage with these tools in our classes and why so I think that's a a better approach for kind of developing that conversation in an ongoing way rather than do this do that yes no nothing is a binary so that's what I would say >> Sarah, we got John and then John back here. >> I represent an organization that brings a message of awareness and education in the public schools uh about sexual exploitation and bullying. Um we need help in our curriculum how to address the issue that is bringing great harm to our kids. Where do we get help and what can you offer today? >> Yeah. >> So, just for clarification, so do you mean the use of AID fakes, for example, and the way that AI tools are being used um you know to represent images of young folks? >> Yeah. Uh I don't know. I don't know. I mean, it is an emerging area. I think there's a lot of litigation going on. I think again with that kind of research-mindedness to look at your peers, to look at major national organizations, how are they tackling this issue right now? How are they helping with that literacy? What policies are schools having in place for this kind of thing? Are schools even aware this is happening? Probably yes, at this point. Um, I think that's how I would think about it is that researchmindedness, what's out there in terms of tools and then that policy question, the awareness question, there are a lot of layers to it that make that work really difficult right now. That's a a really great question and I'm wondering, you know, with the school districts, especially the districts that are providing, you know, iPads and Chromebooks, what they're using because they they have the um oh, I don't know why it's escaping me right now, but with -ate funding and whatnot, they have to abide by the SIPA rules, right? The Children's Internet Protection Act. Um, and a lot of instit like something like GoG Guardian or Securely. And I'm wondering if if there's any movement from those uh companies to be able to provide any type of guidance or support with uh catching that type of cyber bullying or um any type of the deep fakes or whatnot that are happening. Um, I'm I'm a few years removed from those types of conversations since I'm no longer in the K12 space, but I'm really curious if if any of those types of companies have moved moved the needle when it comes to that because the whole reason why, you know, companies like GoG Guardian or Securely, and though I'm not being paid by them to mention their names, but I'm wondering if they have if they have embraced um any changes within the AI landscape to to address u cyber bullying and and and whatnot. So, I'd be I'd be curious if if you're if you're able to have those types of conversations. >> I don't need Yeah, I don't need microphone. Uh but thank you. Question. AI platforms whether it's chap GPT, Gemini, C-Pilot, whatever else. Are there certain niches or certain industries where one is better than the other? restaurants, retail, trades, education, stuff like that. >> There are um depending on which application you're using, right? Because some of them are application based. So like Adobe has Adobe AI, right? So that is that that is AI built for that environment. And so you're going to see a lot of graphic designers, marketing and whatnot who are going to use that. Um in the science um science and research, I think it's called scribe. uh scribe AI. Um so it's pulling directly from research papers when I was in school I was using perplexity and so that the beauty about perplexity and this was before cop before um uh Gemini and all that was was doing this they were actually providing uh the research papers or the links to the articles and before chat GPT was not doing that and so a lot of students were moving in that direction. So, I think it really depends on what application or um or what purpose. I don't know necessarily about like the restaurant business or whatnot, but I would definitely look into, you know, whatever space you're in because there's probably applications um that have that are that have AI built specifically for that purpose. And so, as I was mentioning earlier, you can't necessarily take an AI that was built for a specific purpose and move it into a different purpose because it it's not trained in that space. And so, you know, if you're if you're trying to use a restaurant AI, but taking a scribe AI that was on research papers, restaurant business and research papers, I don't know there there may be a connection or most likely not, right? Um, unless you're talking about nutrition and health. So, so yes, there's there's probably things that are built better for those spaces and and again, you know, you'll just have to talk to your peers kind of what to Claire was saying earlier, too. Thank you. >> You're welcome. Unless Claire, you want to Would you like to >> have something to say, but No, I'm good. I'm good. May I just color in a little bit um on the education piece to John Larson's question? Uh former educator here. Uh the the statewide students are not allowed to use um technology. Whether that's happening or not, we could have another conversation about that. I think to your point, John, it's also happening up in their bedrooms. It's happening on the school bus. have. So technology is just everywhere and and it's how do we how do we get around it, right? So I can say that school districts definitely um have put the kibash on on using um technology in the classrooms. Um so that's one thing. The other thing I was going to mention to your point about um trades and technology and different things. I had the um opportunity last summer to help with an educator extership over in Minneapolis. So um I was with a group of high school educa middle and high school educators and they went to places like um we went to Metronic, we went to um St. Catherine's, we went to US Bank corporate headquarters. The idea was that teachers should learn what industry is looking for when they're growing these students and graduating these students. What skills do they need to have? Without exception, these large organizations said, "Well, they better come prepared to use large language models, even if Metronic has their own, right?" very proprietary. They better come prepared to use it. And of course, they I wish Vicki was here. The teachers were horrified because they're doing everything they can. Some of them are doing everything they can to keep it out of their classroom. So, I as an educator now working in business thought, "Oh my gosh, this river is vast and wide. We're in the same canoe paddling in exact opposite directions." So, I don't have an answer for that, but I wonder if if you can speak to that because at the college level, you're you're a little bit closer then to those folks working, but I just saw the great divide as recently as last summer where industry said, "You better," and high school's like, "We're not." So, >> yeah, I can say I think in a general way, higher ed has gotten that message pretty loud and clear. Um, and institutionwide, there is an expectation that students do need to develop these skills. I think where the m the rubber meets the road really is in the classroom. And I think what helps students is a diversity of approaches to AI in the classroom. So it helps them develop not just a basic understanding of AI, but what AI usage might look like in this context versus that context and also including maybe some classes that don't have any AI in them. I think really it's that kind of class level of engagement with these tools that's going to be the most powerful for folks. Um, and there's a there's a variety of levels of interest and skill across faculty and those of us who work in leadership and administration, too. So, I think actually a strength of what we do is that diversity of approaches. Um, and I think that's really what students will benefit the most from, not sort of a single approach to it. So, my sense is that higher ed kind of as an industry has gotten that message loud and clear. Um, and what it looks like really varies and I think that's actually a strength. Does that make sense? >> Yeah. So I oversee our human resources program. We have a two-year associate degree kind of our approach and that has been students take a lot of HR classes and then at the end they have an HR capstone. So in the HR capstone once our students have had an opportunity to develop their HR expertise in that capstone is when we're saying okay let's try try try using some of these tools and so if they're presenting you with a job description or with a performance improvement plan as a class let's evaluate the kind of content we're getting um is it how is it useful? So we're kind of that's how we've done it in that specific program is to teach people how to use it more towards the end of their program. Not that it always works that way, but that's been an approach. This may be a dumb question, but if AI doesn't know the answer, is it trained to say, "I don't know." Or is it trained to give you back? So yes, it w it will say something in response. I think there have been some attempts by yeah by some companies to say um we won't respond to these kinds of questions. So, if you ask it questions that it perceives to be biased in some way, it might say, "I'm sorry, I don't answer those kinds of questions." Um, and there are some challenges with that. But yeah, it it's I don't want to give too much personality to these tools because they are not human, but their goal is to please you and to give you something that sounds plausible. So, even if you ask them, you know, hey, is this correct? or oh, you got it wrong. It will say, oh, I'm so sorry. You're right. So, it there is no guarantee that um so it won't say I don't know unless it's been built hardcoded into its programming kind of. >> Yeah. >> Yeah. I was um I wrote something recently and it uh there it did provide me a response saying that it could not provide me something. could not provide me feedback because it was not able to because of the content. Um but there was also I believe a case where um a a person had killed themsself because of a conversation that it had with an AI. I think it was chat GBT. So it it unfortunately will always have a response. >> So to some to obviously to someone's detriment. I don't know if >> I was going to share a funny story which is my husband really loves soccer and he does like these like fantasy football kind of pick your favorite players and so he had asked chat GPT basically help me pick my players because he wants to win this league and it kept saying yes I'm working on that I'm working on that spreadsheet and I was like Nate I don't think it's going to happen and it went on for about three days where he kept saying I just I just need more time I'm still working on this I just need a little more time and then he got a little salty with it and finally just said like can you or can you not do this but it took three days of prompting where it kept saying I need more time I'm going to figure this out for you so it will lie to you I'm just going to there is a >> you never got a list >> no because it know it can't do it >> oh my god >> yeah so anyway >> so you can stump it with spell check though my spelling is so bad it can't it doesn't >> that's the test >> that's where it's really smart but also not and then back. >> So why um there's a couple questions. So I work in marketing and I use chat GPT to respond to reviews many different levels but in my like in the evening I watch things like you know the internet made all these millionaires right in chat GPT AI is going to make billionaires. How and how do we do this? >> I'm hearing if I knew the answer to that, I would not be sitting here on this panel is what I'm is what the body language is telling me. >> I I got an answer to that. I mean, I Okay. Yeah. If I knew the answer to that, I I I appreciate being here with y'all, but I might not be here with y'all. I think some of it requires your tolerance for doing shady stuff. So, you know, I think >> I'm definitely >> in I can't help you with that, but things like intellectual property, you know, the environmental impact. So, depending on your tolerance for those kinds of things, I think that is is part of it. Yeah. Sorry, I don't think any of us have an answer to that. Like, how do we become billionaires with AI? Unless you identify a problem that AI can solve and do well. And I want to say the main I want to say software or classes is this cursive and it's like you're going to be the smartest in 20 days and then I go and I AI chat that or you know and they're like no this is not even a good course. It's you can't get your money back. You people are always teaching and I I feel like that is probably where where there's a lot of self-made millionaires like I will teach you how to sell 10 homes in one day. I I better I better go to that class. It's only $300 and you got all my people across the world doing so. I just didn't know if you had like some kind of success stories or you know I know kids can build a website in a day not even five minutes like any anything kind of business fun. >> I don't I never have any fun so I don't really know. I don't know. Maybe other folks have ideas. I don't know. >> Yeah. Go ahead. Well, I think the thing that comes to mind for me is there's some case studies that we look at in our coursework and where a lot of people are trying to find efficiency is with eliminating jobs and replacing them with AI support. And in some of those cases, what we're seeing is that in the short run, you might save money, right? Like I could say, do I really need a marketing assistant? When instead I could just feed information into chat GPT and have it answer. And what ends up happening is when you erode the expertise that exists in your business very often that's hard to sustain. So in the short run there's some organizations that have made a lot of money kind of doing that. So maybe where you would have need needed let's say 40 employees you have two or three people running >> that. >> Yeah. And then and in the end in the end it doesn't work out too well. At least that has been some of the cases that we've looked at in our classes where we're thinking about it. Um, but you probably I mean could attempt something like that. Yeah. >> Well, I think a lot of that wealth is coming from the stock market. Back when you bought Apple and that when you first came >> they said there's three big players that's open >> open AI. All right. Last few questions here. >> I have a question. How are you guiding your students um as they are entering the job market or are searching for jobs with regards to the technology and resumes and trying to get an interview? And some are I'm hearing students telling me they're getting interviewed by AI and they want to make more of a personal um impact. So, how are you addressing that with your students? Well, that's an interesting question because I think on the one hand, we can use tools like AI to help create the most polished resume you've ever seen, right? Um, and so it's important to know how to use it, but I think what we're leaning into is networking and relationship building, right? So, it's important you have your LinkedIn profile, but being a part of a group like this, I think, is more critical than ever. And especially for students who have maybe experienced COVID through high school and they are really into their device and just talking to people and meeting new people is a skill that's harder than it's ever been. I think those are some of the things that we're practicing and being really thoughtful about in the classroom and how do we build those things into our program and offer more supports around things like networking because it's always been hard like to come into this room and not know anybody and you know just kind of put yourself out there. That's a skill in of itself. But I think those are things that we're working on and at the college too to think about how are we providing more support for some of that stuff. So it's a little counterintuitive right we need to know the technology but the technology is kind of putting an emphasis on relationship building and communication skills like never before. >> Mindy, can I ask a follow-up question to that? >> So, given workforce stuff, um, is it appropriate for a c a candidate to call, let's say, an organization, Anderson, Windows, Regents, Hospital, whatever, and just ask the general question, um, do you vet candidates resumes using AI or not? Is that an appropriate question to ask? I mean, I think they could, but I would say from my HR perspective, you should just assume that they are, right? Like, so in hiring and recruiting, and I think if I had the exact data, I hate citing data, that's not how my brain works, but it's something close to like 90% of employers now are using those kinds, regions hospital, and I was shocked to find out. I thought, well, I wonder if it's a question, you know, if you've got these young adults that are looking for job, right? looking for jobs and you know you could say lie and say my name is Bill for instance >> and are you using are you using AI to vet your candidates resumes I don't know >> yeah I think it's a I think that's an interesting thing to reflect on right and so as a faculty member what I'm trying to get them to have is the skill set to do both right so that they've got a a resume that is scannable that would kind of make it through some of that vetting but then also know if they're not using those kind kind of tools. What does that mean for how they present themselves to? But I like that. I like that. That's a good idea. >> And I would say that AI does help get through some pieces of it because you have to be able to meet the minimum qualifications. So for for uh St. Paul College when we post positions, uh applicants have to meet minimum qualifications before they can get, you know, potential interview, right? So AI can help with that vetting process. And so to your point, if there is an I believe it's called ATS uh applicant tracking system, so a that will helps people get through that first part and I would say there's probably a bigger reliance or a heavier reliance on ATS for entry level positions than there are for the higher level positions I would believe. Now I know that we don't use um ATS for our applications either. Um, so but then I would also say too just because you get the just because you meet minimum qualifications doesn't necessarily mean that you're going to do a really great interview. So you do need to have you do need to have the skills. You have to be truthful, right? You have you have to have the skills, but you also have to have the personality and be able to answer the questions effectively when you actually get the interview. So AI is great in some ways, but you also have to be able to be a human and and do the interview. >> Although you could use AI to help you practice interviewing. >> You could. >> That's all. >> Microphone. Okay. >> Well, I'm just um I have a question about like protecting our youth, especially our young women um with you know Tik Tok, right? of all the beauty videos and all of the other videos that are out there and the AI taking their image and and you know putting it out there. So I mean is there any advice on protecting our young women are you know our children maybe not high school of age but you know post high school preol um you know they're they're doing their tutorials. I mean are do should we be reaching out to Tik Tok to you know those platforms to help get them down once their image has been sold type of thing. And I think that's really being it's a huge thing right now especially among young specific. >> I don't know if any anyone has that subject matter expertise. >> Uh but but definitely we always have thoughts. >> Yeah. I mean there certainly there are legislative routes to change potentially. You know, I don't know how responsive corporations are likely to be to concerns. Um, I personally, my opinion is I'm not super confident that that is an especially effective path, but I think in relationship with young folks, um, that the scholarship that I've seen really talks about the importance of, you know, helping parents understand how to navigate those conversations, how to have that ongoing, like we were saying, ongoing conversation with kids about, hey, how are you using social media? What's going on there? I mean, some of you who have older kids might be better suited to speak to that question, but the scholarship really suggests that, you know, parents knowledge of what their kids are up to, you know, thinking about when the right time to introduce social media is in a kid's life. Um, and really building those literacy skills with knowing what happens with your imagery. Um, and then maybe the legal system. I mean, I don't always have a lot of hope about those processes, but it is a tool. So that's that's what I know. But yeah, I'm not an expert in this. >> Kind of to speak to your point about becoming the quick billionaire and millionaire. It's all about the likes, >> right? So if we weren't trying to be the quick millionaire billionaire, we wouldn't be posting our makeup videos, our quick 30 secondond dance videos, our show us our butt videos because that's really what they are, right? Uh and I have a, you know, she's now 21 and I have a 23 year old. So, uh, boy. And so, it's it is all about the likes. It's about, you know, at what point do we start making money off of the pictures that we post and whether or not they are meant to be sexy videos or just like, hey, look at my new makeup tutorial video. They're posting videos to to just, you know, how many thumbs up, how many hearts are we getting? Uh, and so if we if we start if we stop putting the emphasis on how many likes we get and stop making it a competition with our friends, then we we reduce the risk. Granted, we need to we need to maybe help people understand, help our young children understand that >> that there are risks >> that there are risks because, you know, there's always going to be, you know, there's always that internet bravado. There's always going to be the I'm looking at you, my friend, in the brown and the and the blue shirt. You know, you're always going to have the cyber bullies who are going to take that video. You're always going to have the the the deep fakes because now you can do it with your voice. Now you can do it with your image. They're they're always going to have that. But the only re the only way that they get it is if you post it and and I can tell you there was a when I was um when I was in high school, now this was back in the 80s. I was in high school. I was super stupid. I was sitting in an airplane with somebody who was sitting next to me and I happened to say my name was Sonia. I go to Claremont High School and this person came to find me. Now my last name is Zuker Z. Somehow he made it into the admissions office, found the yearbook, found me, went to my classroom. So this was back when there was no social media. and this person creeped on me. So, where there's a will, there's a way. So, now it's just easier to have that way. >> Sarah, we're going to do two more questions. >> Two more questions. All right, >> I got another one. >> Okay. >> How about how do parents get educated on how to deal with the kids, especially in the younger years? because I I'm a firm believer that a lot of challenges in life for kids start at home and I think parents need to take responsibility as they're raising their kids on on what's right, what's wrong, you know, whatever it is. Any any suggestions on where parents can go to get information on how to handle this >> AI 101? Aren't there aren't there parent um like school school? >> There might be school-based resources, but there are national organizations too that have some basic literacy information. I think common sense media is one that comes to mind that folks might have heard of. I'm not sure how much they're doing specifically with AI and AI generated content, but I'm sure there are national organizations with, you know, educational tools. I think you know the challenge is um that parents are coming from so many different contexts with different levels of literacy and time to engage with their kids. So how you address parents as a whole in a way that allows them to access those things that I don't know. Um I will just say for myself personally I have an 8-year-old and we are pretty strict about what our kid has access to right now and will have access to until they're like I don't know 30. So, so you know, but we explain to our kid why, right? They don't use YouTube. They have no access to YouTube, YouTube kids. And, you know, I'll be very transparent with them about why. And that why piece is really essential that parents understand the why. And not just, it's kind of like, you know, college age kids, too. The why of it. It's not just that I'm mean. Yeah. Yeah. That's not going to fly. That's not going to fly with my kids specifically. Bless them. Um, but you know, to have enough literacy to be able to say the why, that is the trick. And then have a relationship with your kid, that ongoing relationship to be able to say, you know, this is what I'm setting up for you and this is not, and this is why. There's no one-sizefits-all for that. Um, for a whole bunch of reasons, but that's how I think of it, you know, and and this is not meant to be disrespectful to any parent, but parents need to be okay with being parents. And I think with this generation and I and I look at my kids too, they want to be friends. Parents want to be friends with their kids. And so at what point are we going to be parents to our kids? Um you know, the way I was raised, the way that I raised my kids as a single parent, we h you know, we had to lay down the rules. Um, and so they didn't get their I made the mistake because I was a single parent of, you know, providing a a a phone to my my son and then learned that he was looking up videos for a friend of his and they were not good videos. And so I learned pretty quickly that while I thought I was being the responsible parent, that doesn't necessarily mean that my children were f, you know, weren't um being peer pressured into doing stupid things on on the phone, too. So, you know, I think that we have to have those conversations and it's it's really okay for parents to say no. And I think that's the other thing, too, is that as we're as you know, I always told my children that I'm going to be your parent. I'm not your friend. I'm your mom. And I think that's that's the other maybe fundamental thing that our parents need to know now as they're raising children, especially in this really scary time of the internet. >> Last question. Do we have a last question? Oh, yes. All right. >> Thank you. Thank you. Um, we were talking a little bit about reports and how AI um, reviews uh, resumes and applications. Um, and I just wanted to put that for a second. And I do work primarily with older adults. And some of our older adults are semi-retired or looking for that last 10-year job. Um, or, you know, or going back into the workforce, thinking they were retired, but now they're not. Um, but where can I access where can the community access resources for understanding how AI is used in the hiring practice for older adults who maybe don't have as much familiarity with that affecting their practices? I think it's a good it's a good opportunity for chambers like this one and others to pull together resident experts and have those um have those conversations, set up panels, even set up a single person to come to community thread and talk through what some of that is. I'm tell I'm not kidding when I say I get pinged every single week about how we are going to get this information out and everybody needs something different. So I would say in in um you know closing here uh that we should probably put out an exit survey that basically says what did you learn today? What what are we missing? What more needs to come? You know John was he's with Anderson so that's a manufacturing like what lanes what part of workforce do we need to tap into? That'd be a good place to stay that we can aggregate that that information but with it changing on the dime. I mean we could run this quarterly >> and I think we would we would all learn something. So, good question and you and I can connect. >> Yeah, that'd be great. >> And I think although AI is new, you know, sort of introduction of technology to folks that are unfamiliar with it and are thinking like I don't need to know that, but maybe they do. You know, I'll I'll answer this question that, you know, 20 years ago at our college, we were one of the first institutions in the community college network to go online and require applicants to apply through our online portal. Like we said, we are no longer accepting paper. It's a matter of efficiency. Um, but what we did as a part of that is we modified our admissions office and had 20 computers installed. And so if someone needed help, we said, "You can't do it by yourself. You don't have a computer. Come on in." And we would handhold them through the process. And sometimes it was like, "It's asking for your first name. Type it in. Let's type let's find the letters." Um, and so I think that to a certain degree with anything where you have folks that are really new to a concept and trying to learn, um, offering hands-on support sessions to help them navigate it is a really important element. >> Public library, too. >> Public library. Yeah. >> Anyone else wanted to answer that question? All right. Thank you very much for having us today. Thank you for your questions.