The Center for Public Sector AI: 2024 Generative AI Update for Education Leaders

Hello and welcome to today's webcast, the Center for Public Sector AI 2024 Generative AI Update for Education Leaders. I'm Brian Cohen, Vice President for the Center for Digital Education, and I'm excited to serve as the host for today's webcast. Thanks for joining us. We're gonna have a great session over the next 30 minutes. But before we dive into the content, we have a few housekeeping notes for you. A copy of today's recording will be sent to all registrants within 48 hours. Go ahead and use that recording for your reference, or feel free to pass it along to your colleagues. As you will see, this webcast is meant to be interactive, and you can be— you can participate by submitting your questions using the Q&A function at the bottom of the screen. And finally, don't wait to send in your questions. Go ahead and submit them as they come up. and we'll address as many of them as possible during the Q&A session. Okay, let's get started. Joining me today to discuss this topic is the president of eRepublic, Dustin Haisler. Over the next 30 minutes, Dustin's going to share some great insights and a really great announcement. So before we— without any further ado, over to you, Dustin.

All right, well, thanks, Brian. I appreciate it, and great to have the opportunity to talk a little bit about what's happening in the generative AI landscape as it relates to education. That big announcement that Brian mentioned is actually we rolled out our Center for Public Sector AI's official web presence. For those of you that don't recall, in December we announced the emergence of the Center for Public Sector AI. This was the first research institution really geared at trying to help make sense of this unique moment for state and local government and education leaders across the country. So definitely plug into that. There's tons of resources. Market intelligence, all of the webinars that we're doing are all there as well. So highly encourage that you check that out. But let's go ahead and dig into what's happening with AI across the country. And I thought I'd start first with just a quick AI refresher on the landscape. So when we think about kind of artificial intelligence as a macro category, the lion's share of attention today has really been on generative AI, which is a subset deep learning, which is a subset of machine learning, which is a subset of the AI category. Now you'll notice from a regulatory standpoint, this gets into interesting, you know, territory. If your districts are passing policies on AI as a category, are you really trying to regulate generative AI solutions, or are you trying to regulate AI as a macro category? What you need to know, and really the best way to kind of think about generative AI, is this is AI where human is in the loop. This is AI that enables and augments humans to create. It's AI that's really focused on the creation of net new based on what it's been trained on. So even as you think about GPT, generative pre-trained transformers, basically what you need to know about that is it's just pre-trained AIs that you can leverage that take that base of knowledge and create new things, whether it's code, whether it's text, whether it's images, video, all kinds of new things that are emerging. Now we've seen a, a quick evolution of AI from standalone AI, which is where you go to a specific place to do something like ChatGPT or Google's Gemini or Copilot or other things. Now into more embedded AI where these things are starting to work their way into the productivity apps that we use every day. And, you know, providing ways even at the operating system level where we can interface with generative AI prompts. That's also rapidly starting to change as we have seen a lot of announcements over the AIs that are starting to work their way into our phones, right? As we think about the evolution of the voice assistants we have on our phones, you know, will we see, you know, generative AI there this year? Yes. And so expect to see that AI also start to abstract the way that we access some of the other solutions? And then this gets us into this other interesting territory of, of, you know, autonomous AI or semi-autonomous AI that can operate on the macro goal that we give it and go and, and find and, and work on relevant information on our behalf. So a wild change of different things, and all of these domains are having so much innovation and disruption. This is also creating the emergence of a new aspect of our tech stack as we think about the traditional environment of technology within our districts. Now that infrastructure layer has an AI application to it as well. We have to think about the underlying AI infrastructure that we're using. We also have to think about the models that are built on top of that. Are they open source? Are they proprietary? Are they homegrown? Are they models that we're leveraging commercial application or technology for? And then the applications that are built on top of that. Now, each of these different things, just like a software as a service technology, can have a high degree of variance, right? You might be using SaaS software today That has its own AI tech stack. You might want to do something for your own district. So there's going to be a little bit of fragmentation here, and we need to think about the orchestration of all of this, especially from a records retention standpoint. And as we think about new regulations that will emerge with training and compliance with bias and, you know, plagiarism and other types of things, there's a lot to consider here that we'll have to confront. Now, where is this headed? You know, as we've looked at education institutions across the country, You know, we've seen 3 different domains very rapidly emerge here: people augmentation, knowledge augmentation, and experience augmentation. So, you know, in essence, people augmentation. Every one of your employees and every one of your students will have the equivalent of a personal assistant in their pocket. That's going to drive a ton of productivity gains for those districts that embrace it. There's going to be data-driven insights to aid in the back-office administration, even the instruction dynamics. There's going to be project management and other types of workflows that are all streamlined. and even skill set augmentation where there may be gaps today, you know, individuals will have a ton of capabilities that can augment them in their pocket and with them. Now, from a knowledge standpoint, a lot of institutional knowledge leaves the building as we have, you know, fluxes in our workforce. And as people retire, we're gonna start to see education knowledge augmented through generative AI systems that allow not just the automated and autonomous capture of this, but the accessibility of it. Think about all the different things that we have, all the data streams that we have within education. This is going to be accessible in a conversational way, right? I don't have to have a data scientist that is on staff that's helping me navigate this. I'll be able to start to leverage some of these conversational prompts to deal with and interact with the information that I have. Now, there's, of course, a lot of things that go into that that have to be confronted, like AI hallucinations and bias and even PII and how we embed and protect that information. Is our data being used to train someone else's model? A lot Lots of things to confront there, but this is going to provide a renaissance of how we access information within our districts across the country. And then experience— the way that people interface with us, the way that people experience what we provide will change. There's going to be not just the hyper-personalized learning that's talked about, but the personalization of internal and external communications, greater transparency in what we're providing, and accessibility into the output of that, both for teachers, for parents, for the back office administrators. And then all of this will increasingly become abstracted as we think about where people find information. You know, are they gonna go to your district website and analyze everything that you have there? Are they just gonna use a generative AI solution that's closest to them to analyze and compare districts and scores and methodologies? It's gonna be that. And so we need to start to think about how do we optimize for that? How do we optimize our content for that? And how do we enter a world where maybe our website is not the front door to individuals learning about us going forward. So a lot to kind of confront there. Now I want to spend the majority of our time digging into just some updates, and the best place to get this is through our beat with the Center for Digital Education. We are covering what is happening with AI and a lot of other domains as it relates to K-12 and higher education. We're also curating this on our Center for Public Sector AI's website, so definitely check that out as well. Now if you've got questions as I roll through this, definitely use that Q&A feature, and Brian and I will try to tackle those at the tail end of this. But lots are happening here. Now we've seen that, that cautious optimism associated with generative AI really start to transition into a more proactive stance with how districts are approaching this. We've seen, you know, new guidance that's emerged that's kind of moved from the interim guidance of the past to like, this is a real thing. Now we're starting to implement it. We saw at the K through 12 level, Washington State created new updated resources that, you know, provide some guidance for their respective local jurisdictions on practical implementations, terms, and facts, trying to take an enterprise approach approach to how this gets rolled out across the state, even at that local district level. Now, we've also seen a lot of new hats that have been brought into this, right? As we think about the, the people that are responsible for AI, who in your district is responsible for the implementation of AI? Is it your CIO? Is it your CTO? Or is it a vice dean of AI? And so this is another dynamic that we're watching very closely. Where does AI sit within the organization? And how will education institutions grapple with that and scale that across the country? So definitely, keep in mind what— and keep a lookout for what is happening there. We've also seen, of course, higher education leaning into this fully and trying to leverage this as a differentiator in terms of what they provide to their students. So a lot of the computational computer science-oriented aspects of higher education are evolving into AI and data science. And so no surprise there. But this is also becoming an area of focus just from a curriculum standpoint. Now, what really fascinated me is when we started looking at some early research here around the use of AI in districts themselves. And so we know, like, from a back office standpoint, there is limited professional development that's been offered there because it's so new. And also, as we think about the actual individuals that are doing the instruction, as you can see from this recent survey, you know, over 7 out of 10 teachers with this national survey had not received any type of professional development on AI use. And I think this is indicative across the board of all the public sector. There's a huge need not just to implement policy guidance, guardrails, and the, you know, other things that go along with that, but to also find ways to ensure that we're focusing on reskilling, upskilling, and bringing everyone along so they understand how to use it. There's a lot of new use cases that are emerging as well. We've seen the AI dynamics now starting to work their way into vision recognition around AI-powered surveillance systems and campuses. But even this has some challenges. There's groups that are out there that are trying to take some of these emerging technologies and ensure that there's guardrails that are put in place on them. So, you know, as you think about enterprise rollouts of artificial artificial intelligence and generative artificial intelligence, you want to make sure that you watch that regulatory landscape closely because there are counterarguments to all of these different adoption points that you're going to want to follow. And some states are taking action on that. We've seen a lot of proactive states. I mentioned Washington. This is Ohio. They released an AI toolkit for their schools compiled by the state. So again, you know, centralized approaches to trying to, you know, really move on these technologies, incorporate them. Into the administration and instruction of these districts. Now, we saw a lot of things start to change actually with the rise of the generative AI app stores or the GPT stores. So OpenAI announced the ability for anyone to create their own GPT. And again, a GPT is just a pre-trained set of language that you've optimized based on its body of knowledge that you can use for your own specific purposes. And so OpenAI released their own version of that. We've seen Vertex AI kind of come out with their own version of that that was announced this week. But it's been interesting to follow the number of districts that are leaning into that and creating their own pre-trained GPTs for accessing district data. Maybe it's for teachers, maybe it's for administrators. These are just some of the public ones in a quick search of a school district. You can create internal versions, you can create your own version for you, or you can create a public GPT that again is pre-trained on whatever function you want to optimize for. And this is creating a flurry of like what we call the App Store moment, a flurry of excitement and also fragmentation with AI. as districts are trying to understand how to implement this and how this fits into the overarching experience. Now, we've also seen this rolled out as we look at the higher education level. In-house AI is here. We've seen 2 university campuses, UC Irvine and UC San Diego, rolling out their own campus-wide web-based AI programs for staff and faculty, again, on the back office side to help streamline that. Students are soon to follow. Ohio educators have an AI known as Ally in the classroom as well. So this is something that's hit a couple of those school districts there as They're testing these types of pre-trained bots that can basically help be a resource for teachers and students. And then, you know, the ZotGPT is another example, as I mentioned, with UC Irvine that's built on that custom kind of generative AI platform. So again, a lot of movement here, a lot of focus, and it's gone from just let's build some interim guidance and test the waters to how do we implement this and what is an enterprise approach to that? Now, there are some challenges that you want to be aware of that we're noticing here as we look at you know, K through 12 and higher education. And so I want to spend a moment just talking about those. Number one, the way you procure these technologies will have to be updated. As you think about the way that AI is actually adopted, if you're buying a new solution that leverages AI or generative AI solutions, you're going to want to make sure that that's disclosed within procurements. You're going to want to make sure that you ask the vendor community and the industry community to be transparent about that because you're going to start to get challenged on that and you're going to have to produce records associated with that. And you're going to want to make sure that you have those things together. Districts will also need to make sure that, you know, the legal indemnification regulations are put in place from a procurement standpoint. If you're creating images using an off-the-shelf consumer generative AI, you want to make sure that not only you disclose the use of that, but that you're indemnified for use of that image for any future lawsuits that may emerge. And trust me, there's a lot that are happening right now with artists and other individuals that have been used as a basis of training some of these models. Now, as we kind of continue on our path of emerging challenges, is we also have BYOD. So for those of you that remember bring your own device, that has come and gone. We admitted defeat. Devices are here to stay. But now as we think about the fact that consumers have access to these models, so it could be a teacher, it could be a parent, it could be a student, it could be an administrator, they're leveraging AI as an aspect of doing their jobs. And so how do we think about records management associated with that? How do we maintain privacy and security associated with that? So there's a lot of dynamics here that are going to have to be confronted more than just policy. Privacy policies and disclosures will have to be updated as well. You're gonna make sure you think about things like how are you using someone's data and are you training your own system to be optimized on existing data that you have? Is it going into someone else's system to optimize their language model? So there's a lot of new areas that are starting to enter the picture here. The federal government has issued a lot of guidance for federal agencies on the need for training and the need for disclosure, and we expect to see more of this to hit the state level, and it will have direct implications for higher education institutions and K-12 institutions in their states over the course of 2024 and 2025. And then, as I mentioned before, AI literacy, reskilling, upskilling, all these things are incredibly important for us to confront going forward. We need to think about professional development in light of these things. It's not just a policy that comes out or example guidelines. We need to go beyond that. Now, we know that procurement and ethical procurement will continue to be a challenge, and there's a lot of new standards that are We also know that the digital divide is another dynamic here. As we mentioned, you know, the back office implications of reskilling and upskilling employees. There's also the same type of challenge we need to think about with leaving people behind from a digital divide standpoint. We don't want this to be the next digital divide. We want to make sure that AI is inclusive and that our approaches from an enterprise standpoint within our districts balance that equity piece in everything that we do. And so this will become an important part because as you think about a lot of these systems, they require high-speed connectivity. They require access through a device. And so how do we ensure that there are offline ways to access, there are other means and mechanisms for people to plug into these systems and gain value? And then there's the cybersecurity implications of this. We talked about the new tech stack at the beginning. We are seeing the cybersecurity threat landscape light up across K-12 and higher education. And why is that? Well, a lot of those bad emails with bad spelling and capitalization issues, before are now starting to look really polished because there is a dark web of capabilities that are leveraging the innovations happening with generative AI, but they've created their own dark web versions. Case in point, WormGPT and FraudGPT. These are 2 examples that are on the dark web that threat actors are using against districts. And so they're building very sophisticated phishing campaigns, and they're finding ways to create deepfake audio recordings. And, you know, the level of sophistication for these attacks, even for local districts, is incredible. And so this is going to require a new form of training, a new offense from a cybersecurity standpoint as we think about how do we leverage AI to combat some of the new threats that are emerging. But we will see a lot of activity. And again, all this feeds into a ransomware-type attack or malware injection or variety of other types of things. But their front end of that now is becoming a lot more sophisticated and becoming infinitely scalable through web architectures and infrastructures. But I think this is a unique moment. I mean, as we look all of this, you know, the federal government has a good stance here that I agree with. You know, this is not an opportunity to sit out. So do not sit out on AI. And I think this is case in point true with the K-12 districts and the higher education institutions that we've seen. Those that have restricted it have ultimately set themselves back. They're playing catch-up. They're trying to figure out how to get with some of these technologies. Their staff are already using some of these technologies. So these are becoming more prominent across the country in use and adoption. And so this is an opportunity to lean in in a safe and responsible way, understand where things are heading, learn from some of the others that have kind of gone before you, and find ways to figure out how to put the right guardrails in place that are contextual to your district type. So you definitely want to make sure that you follow that and kind of keep up with that. Now, we are tracking all of this through the Center for Public Sector AI. So, you know, highly encourage you to plug into that. We have a weekly newsletter that goes out there. But I want to spend the rest of our time by bringing Brian back on. And I know I gave you a fire hose of kind of some of the initial updates that we've seen, and we will have some others that are coming in the future, both on the government side and the education side. But we want to answer your questions and hear from you as well. So I'll invite Brian back. So hey, Brian, welcome back. Let's, let's kind of tackle some of these questions that we're getting. I'm also going to ask our audience a few questions of our own just so that we can better understand how they are approaching these topics. So if you have questions, definitely, you know, shoot those in through the Q&A feature at the bottom of your screen. We're gonna ask you a few questions too. We'll share those results just to see, you know, where you're at across the country from an institution type and how you're leveraging generative AI today. So, you know, Brian, you know, as a former higher ed CIO of one of the largest districts in the country, tell me a little bit about from your perspective, you know, what should education IT leaders, you know, be thinking about today as they navigate this crazy period of change?

Crazy period of change is right, Dustin. You know, I haven't seen this much action in IT and education in a long time. So, you know, I spent nearly 20 years with the City University of New York, and there's a lot going on in education, both at that K through 12 and higher ed level. But, you know, what I've seen when you get, you know, is really the focus is that everyone is thinking about AI and how they're— how they need to get ready for AI, you know. They, they, they're, they're focused on, you know, how is it going to work in their universities and how do they find that right balance between being innovative and on the leading edge or bleeding edge, as well as being, you know, maintaining that risk level and that level of security. You know, I think it's super important that, you know, they continue to focus on the new technology and the advances that are coming out. I think that, you know, the CIOs and their staff as well as the people that they're supporting really need to be, you know, in spite of the fact that they're swamped with day-to-day work, they need to really be thinking about, you know, how they educate themselves, how they continue to develop their own professional experiences and see how these tools like that you've— that you brought up today are gonna fit into their environment. And, you know, I used to use my— with my team, we used webinars and conferences and events like this to really help educate ourselves. So it's really important that you find that time to carve out of, you know, in your week or in your day to do this. And, you know, I would also say that, you know, it's about staying ahead of the curve. These CIOs are— they're trying to be ahead, and they're trying to be 2 steps ahead of the academic and the administrative community so that they can be ready to react rather than sort of, you know, sitting there and saying, we're not ready here, we can't do it here.

Yeah, great advice. So I'll take this next question. Yes, we will be sending a copy of all the slides out from today. So within the next 48 hours, you'll all get a copy of the recording from today's event as well as the slide deck. So feel free to use that. You know, also, as you think about your district, if it's helpful for Brian and I to talk to your team, to talk about what's happening, to kind of share those national perspectives, we're happy to kind of do that as well. We want to make sure that we're a resource for all of you. All right, we've got a question in the chat here around, can we talk about where higher education administrators can begin and in terms of learning about how to mobilize AI for work efficiencies? I could take a first crack at that, Brian, and I'd love your thoughts.

Yeah.

I think this is one where, you know, one, you know, creating a time and a place and a governance group that can talk about these things will be incredibly important. Most of the time, you know, we think about technology updates once a year, right? We meet with our IT team, we'll look at how the, you know, world has changed, we update our plans. This is moving every month. So at a minimum, you want a quarterly, workgroup that's associated that can talk about how things are changing, what the new regulatory environment is looking like, what the use cases are that are emerging, what they're seeing, right? It could be from them, it could be from their kids, whatever it might be. But you have to create a more often time and place for that conversation to take place. That is the most common approach that we've seen that's really paid dividends is creating a task force or workgroup of type— of some type that can kind of navigate that, have that conversation, adjust policy, provide opportunities for learning. And then you want to start to look at how you bring that new stimuli into the rest of your organization. And so, you know, that's where you want to look at opportunities to diffuse. What are other districts doing? What are other higher education campuses deploying? How can we learn from them? How can we be inspired by their action? How can we partner with them going forward? So you'll find some natural points of connectivity there. But Brian, I'd love to get your perspective on this as well.

You know, Dustin, I agree with you 100%. I think this is a grassroots effort at a lot of higher education institutions. Look, higher ed is known for its collaboration, and I think you can't stop, you know, thinking about how AI is going to be impacted with collaboration, right? You know, you administrators need to talk to everyone, talk to the faculty, talk to the staff, talk to the IT guys, talk to the folks within their organization about how AI could potentially be beneficial to them. But by talking to their own staff and getting their own staff's ideas and perspectives, Nothing is— no idea should be a bad idea. Everything should be explored. We used to do this in my institution. We had blue sky type of meetings with just, you know, the staff to say, hey, if nothing else was in your way, what would you want to do, or how would you want to use that technology? You know, and I think you're right. I think it's the conferences, it's the webinars, it's reading about case studies and where other people are doing things. And I wouldn't sort of just look at just higher education. I think higher ed administrators should look at what the private sector is doing in corporations and businesses because we can learn and we did learn a lot from what they did back then.

Absolutely. And I'm going to share the results from our last poll question around have you used generative AI for work purposes in your district? 71% of you said yes. How often for those of you that have done it? 16% are daily, 26% weekly, 32% monthly, only 18% of you not at all, and 8% are unsure. So, you know, this is definitely mirroring what we're seeing across the country. I mean, I think the wide set of responses for, you know, leveraging this for work purposes, you know, the party trick days of generative AI to write a poem or a rap for you are over. And now people are finding ways to incorporate it into their workflows. So, so definitely, you know, some good insights there. Any surprises for you there, Brian?

You know what I'm surprised at? 21% said that they did not use it in their workplace. You know, it's either because there's a policy against it, or maybe there's a fear about it, or there's a lot, you know, there's a need to know what it can really do for them. But it's almost like asking the question, do you use the internet in your workplace today? I mean, everyone's using the internet. There should be a reason or a value that people can get from using AI. I'm, you know, that was the only surprise for me.

Yeah, great insights. This is a good idea and kind of quasi-question here from Patrick. Is the center planning to host workshops where selected districts and universities provide live demonstrations of their AI products. The, you know, 35,000-foot view is helpful, but also seeing what's on the ground is also good. So any thoughts on that, Brian?

Oh, I, you know, I love that idea. You know, it's— I'm the type of guy who used to sort of like sort of stay at the 100,000-foot level, but I love to get into the weeds. And I think that's exactly what Patrick's suggesting. We should look at sort of ideas in which we can do some case studies and talk about how AI is actually specific AI is how it's improving the particular work that someone's doing in higher education. You know, and I'll even tell, you know, Patrick, if you have some great ideas, or you have some examples that you want to share with us, please send them to me through LinkedIn or through my email. And, you know, those are the types of things that everyone in higher education and even K through 12 would benefit from learning about.

Absolutely. And I think the best way to think about the work that we're trying to do through the Center for Digital Education and the Center for Public Sector AI is we want to be a platform. I mean, Brian and I were both on the other side of the I was in government, he was in education, and we're here because we want to be a platform and a resource. So, you know, we want to provide air cover where we can. We want to provide helpful, you know, information, but we want to create those forums where you can share those things so we can broadcast them with our wider network. So if you've got great use cases, ideas, or other things like that, as Brian said, you know, send them to him, send them to me. We are all about trying to do that because we know the landscape is moving so quickly and you're going to need, you know, action-oriented content and other types of resources to help you navigate what's happening right now.

now.

So last question for you, Brian, I know you were just at COSIN. You know, what's kind of the pulse that you're hearing with AI at some of these larger-scale events across the country? Tell us a little bit about, you know, what's happening out there in the field.

Sure. So, you know, COSIN's focus is really K-12 education. And I'll tell you, the whole conference was buzzing, buzzing with superintendents, super educational edtech people on the academic side, and the administrators, the CIOs were there as well. And the, and the Chief Information Security Officers But boy, everybody, every conversation that I was having was about AI. It's just, it's just huge. You know, they're all thinking about, well, you know, it's cool tech and how do I sort of embed it into the work that I'm doing and how do I learn more about it? They're, they're, they're realizing that it's not just about the technology. While that's really cool, they also need to be responsible in how they're doing it. So they're focused very much on the you know, the privacy issues, the cybersecurity issues, the educational piece about it. You know, it's— you can't expect people to understand or not necessarily use the technology in a safe way unless you in fact provide that level of awareness and professional development. So I think those were some of the big important messages that were coming out of it. And I'll tell you, if I give you some perspectives of the angle How can it support the mission of the district, the superintendents and leaders? They were focused on that. They were focused on how can it improve student outcomes? They were thinking about that. They were asking questions about how does this support the job of teachers, make them more productive, take away some of the minute tasks that they have, like writing comments on report cards. Those things all take time. They were focused on how do they make their district safer? And the real cool one was really how do they leverage AI Not just for physical, not just for the virtual security that we think about, but physical security integrating into their camera systems and making their schools safer for students, safer for the building as well.

Love it. Well, we promised you a 30-minute kind of rapid-fire webinar on the state of what's happening with generative AI across K through 12 and higher education. We hope that we have at least given you a little bit of a fire hose as to some of those trends. We're going to be doing a lot more of these, so keep the ideas flowing. You can also make sure to plug in at the Center for Public Sector AI's website, links on the bottom of your screen. You can scan the QR code, connect with Brian and I on LinkedIn. We'd love to hear from you, stay connected, see what you're doing, and also let us know the challenges that you have. If you're facing difficulties and there's a gap in knowledge, let us know. We want to, we want to be able to confront that. If we can help you by presenting to your district on Zoom or whatever it might be, also let us know. We want to be a resource to all of you. Thanks for carving out 30 minutes. And we look forward to seeing you at one of our future events. Brian, always fun to share the virtual stage with you. We'll see you next time.

Thank you, Dustin. It was great.