Utah is hosting its 2025 AI Summit on December 1-2, calling the event “Utah’s Pro-Human Leadership in the Age of AI.” The gathering will bring together Utah leaders in “government, business, and academia to engage in pivotal conversations about the future of artificial intelligence.” One of the many speakers will be Matt Winters, the Artificial Intelligence Education Specialist at the Utah State Board of Education.
To learn more about AI in the realm of education, I reached out to Matt Winters for a Q&A, which is found below. Read on for his expertise and insights on AI in education policy.
Fairbanks: Many people feel AI in education is a new area, but AI – even generative AI – in education settings has been around for a few years. What lessons have we learned from on-the-ground education practitioners so far?
Winters: Generative AI has been around in a consumer capacity since November of 2022, but the world of AI goes back to the 1950s. AI has been a large component in the background of many educational technology tools for the last couple of decades. We can look at things like Google Translate, Maps, and the value of technologies like those available in standardized testing. However, generative AI is adding a new component to the world of AI in educational technology.
So far, we’re exploring what researchers like to call “emerging practices.” Emerging practices aren’t necessarily set up to be best practices, but we’re starting to see what is happening in the field and what kind of best practices may arise because of them. However, because generative AI is, at this point, less than three years old from a consumer standpoint, it takes time to develop those emerging practices, correlate them, research them, evaluate them, and then find enough of them to become a best practice.
At this point, what we’re learning is that educators are exploring. They’re digging in, and they’re seeing how it works. We’re encouraging them, from a state board perspective, to do that in a safe way: to use approved tools that meet the floor of data privacy and to build it in ways that move past productivity into creativity, where they are able to build something they’ve never been able to build for their classroom, or accessibility, where they can meet the needs of all their students in a variety of contexts.
Through work here at the Utah State Board of Education, we have been able to meet with many educators, see them go through training on artificial intelligence, and we’re starting to see some of these emerging practices start to rise to the top. These are things connected to creativity – building things that you’ve never been able to before – accessibility – helping all learners in the classroom – and honestly, productivity, where it speeds up some of the processes in our classrooms.
Fairbanks: What role should public policy play in AI and education? How might it help or hurt AI’s potential for schools and learning?
Winters: Utah has been really great at supporting policy around education and educational technology throughout the last few decades. Programs like Digital Teaching and Learning here at the Utah State Board of Education, and building the Utah network for internet access through Utah Education Network, have been incredibly invaluable from a policy standpoint to support building out artificial intelligence’s role in education. It’s given us a great baseline for how and when educators are using artificial intelligence and their dispositions around that use.
However, we need to think about how and when policy could be extended to better support not just AI, but other uses of educational technology in classrooms that meet the needs of our students and teachers. One great area that Utah has led on in the last few years is student data privacy. Student data privacy is a deep concern around the entire globe right now, but Utah has taken a leadership position on the issue and how it interacts with a variety of educational technology tools. The main perspective is: how do we keep our students’ and teachers’ data safe, regardless of the tools that they’re using? That has been a fantastic policy move to support not just artificial intelligence, but a variety of educational technology tools.
Fairbanks: We often hear of AI in limited terms, like students cheating or helping teachers create lesson plans. Help us look into the future. What are some exciting, or even less-discussed, potential uses for AI that can support education?
Winters: Artificial intelligence, particularly generative AI, is what many researchers consider a fundamental technology. It’s a lot like electricity; it only gets really interesting once we plug something into it. In the case of electricity, that means light bulbs, laptop computers, and any number of electrical devices. With AI, it gets really interesting when we plug human creativity into it.
In a lot of trainings that we’ve been working on at the Utah State Board of Education, we’ve seen teachers move from productivity, where they’re able to move through basic tasks more quickly and effectively, to creativity, where they’re able to create things that they’ve never been able to do before for their classroom and really unlock some of those amazing things that are sometimes more difficult or time-consuming for our teachers to complete. AI opens up that potential for our teachers and our students as well.
But then, a lot of times, our teachers get towards this idea of accessibility. What can AI do to support all of my learners in my classroom? So yes, we see a lot of lesson plans, and we see a lot of emails, but the reality is that we also see some amazing uses of what we call “constitutional AIs,” where a teacher can pre-program an AI and say, “I want you to interact in these ways, with this kind of scripting, with this kind of methodology – maybe it’s Socratic methodology – and help our students work through a problem.” For instance, it might be building a bot that has a constitution that focuses on teaching students about the formation of the US, with backing documents included in its knowledge base, and from the perspective of a boots-on-the-ground person who lived through the Revolution. All of that could be done really easily in artificial intelligence, particularly in an AI chatbot, and we’re starting to see more widespread use of this in classrooms across the state of Utah.
Fairbanks: What are some of the challenges you see ahead?
Winters: One challenge that we see across artificial intelligence that our teams at the Utah State Board of Education are really paying attention to is making sure that we have a human in the system when it comes to AI use. A worst-case scenario would be that a student uses artificial intelligence to generate the product for their assignment, the teacher uses artificial intelligence to create comments about that product, and the artificial intelligence is the only thing that is learning in that situation. So, we are encouraging our districts and charter schools to really think about when and where decision-making is being taken away or set aside from a human standpoint.
This is called substantive decision-making. We want to make sure that when artificial intelligence is being used, the human in the system is still making all the substantive decisions throughout the learning process. One problematic area that we’re seeing on the horizon is the rise of “agentic AI.” Agentic AI allows users to target specific problems or ideas with artificial intelligence so that the AI takes the decision-making process out of the human’s hands and allows the AI to make decisions for the human. Right now, it’s being used for planning vacations and other everyday items, but when it comes to grading, evaluation, and learning, we want to make sure that students and teachers are not just in the loop, but are making all the key decisions within that loop.
Fairbanks: What’s helped make Utah one of the leaders in this area, and what will keep us in the leadership role?
Winters: Utah has been leading on artificial intelligence because of three major factors.
The first factor is a long-tail investment in innovation. In the early ’90s, our governor asked Utah Education Network to build the backbone of the internet for schools across the state, which has provided free internet access to all Utah schools. That’s innovation on a long-tail scale. In 2016, we saw the launch of the Digital Teaching and Learning Grant through the legislature at the Utah State Board of Education. This has provided grant funding to districts and charter schools to start technology programs, fund coaches, provide professional development, and meet all sorts of professional needs around technology. Those are investments in the ecosystem around educational technology.
The second thing is that we have deep collaboration in the state of Utah to support innovation at all levels. Whether it’s collaborative groups put together by experts in statewide systems or teacher groups at school levels getting together to share expertise, Utah is rife with examples of collaborative learning and engagement in the educational system.
Third, Utah has a deep pool of expertise at all levels of education, particularly around artificial intelligence. Whether it be our Office of Artificial Intelligence Policy, which is part of Utah’s Department of Commerce; our higher education institutions that are supporting responsible AI use and research; or deep investment from industry partners in understanding how and when AI should be deployed, used, and explored in a variety of contexts, Utah has deep expertise that is being willingly shared at all levels.
These three components – deep investment in innovation, collaboration, and shared expertise – are providing the backbone of how Utah is leading on artificial intelligence in education.
End of Q&A
Conclusion
Sutherland Institute has been sharing insights from AI experts in a variety of public policy areas. To learn more, you may want to check out these Defending Ideas podcast episodes on AI in education, AI to improve elections, and AI in healthcare.
