As college students across Texas return to classrooms and lecture halls this month, one website is sure to be in many a bookmarks folder: Chat GPT.
With the seemingly inevitable rise of AI chatbots, will universities be able to use AI to enhance student learning while preventing the worst academic abuses?
Julie Schell, assistant vice provost of academic technology and the director of the Office of Academic Technology at UT Austin, recently helped lead an effort to shape the university’s guidelines around the use of generative AI.
Texas Standard spoke with her about the future of AI in higher education. Listen to the interview above or read the transcript below.
This transcript has been lightly edited for clarity:
Texas Standard: UT recently released new guidelines for the use of generative AI in teaching and learning. How is the university approaching this issue?
Julie Schell: We approach AI with a framework that we call “AI Forward, AI Responsible.” And that really means that we want people to use AI in ways that will foster learning and also have a strong sense of its limitations, and to approach teaching and learning with those limitations in mind. That is the “AI Responsible” side.
AI companies are now marketing directly to college students. How do we balance the excitement around this new technology and the potential for very real academic abuses?
I think it’s really important to help educate our students and our faculty on those limitations. And one of the things that we’ve done here at UT Austin is early on, a year and a half ago, we identified what we call the “big six.” And those are the big six limitations of using generative AI for learning.
And we’ve worked really hard to make sure that our students and our facility understand what those “big six” are, so that when they do engage, they’re able to use those muscles and their own integrity to make efficacious decisions about how they’re adopting these tools for their own learning experiences.
One of the big concerns is that, with access to AI, students are going to cheat all over. How do we prevent that? Does this mean a return to handwritten exams, or is it something else?
Definitely academic integrity is always a concern when you have new technology come to the surface. And one of the things that I think about a lot is that preventing cheating is a policy goal, and learning is a design goal.
And so I think we’re in a moment right now where we really have an opportunity to kind of fix some of the things that maybe have been off in education for the past little bit around how we do assessment in particular. We have to rethink what meaningful and authentic assessment looks like for our classrooms.
» MORE: Recent college grads have new competition for jobs: artificial intelligence
It sounds like it’s not that UT is discouraging AI use in all forms. In fact, there’s some encouragement of using it wisely. Can you talk a little about that?
Absolutely, I would say we’re not discouraging use, but we are discouraging irresponsible use.
And we think that use is important, particularly for where we are in the moment in the world right now with AI, but responsible use is equally important, and it’s part of being intentional and helping build up that ethical muscle that we have to help us understand, well, I can use it, but should I?
What makes you most excited about AI’s potential uses within academia?
That’s part of one of the things I’m most excited about for this fall. So in addition to having frameworks, we’ve developed an AI tutor called UTSage. And UTSage is designed to address those limitations, those big six limitations like privacy and security.
And another one of the limitations of using AI that we worry about, in addition to academic integrity, is something called cognitive offloading. And that means when you offload something to a machine, you don’t want to just let that space that you freed up atrophy. You want to fill it up with additional critical thinking activities and things like that.
So one of the things that we’ve done is we’ve developed an AI tutor called UT Sage, and faculty members can train these tutors to teach any variety of topics that they’d like to teach. And Sage has a science of learning backend to it that helps with this.
Instead of just giving the student a direct answer if they ask a particular question, they’ll help the student engage in more reflection about that particular question and try to extend that learning and that critical thinking and that discernment in sort of a workflow for the students.
So I think that’s something that I’m really excited about, the ways in which we can train these technologies to serve as tutors for our students to help extend their learning beyond the time that we get to spend with them in the classroom.
» GET MORE NEWS FROM AROUND THE STATE: Sign up for Texas Standard’s weekly newsletters
You’re teaching a class this semester yourself. How are you going to talk about using AI with your students?
So what I like to do in my own classroom is scaffold the use of AI for the students. And so we talk a lot about responsible use at the beginning of the semester.
My class is structured around three projects. And in the first project, students can elect whether they want to use AI or not. In the second project, I actively encourage them and teach them how to incorporate AI into their projects. And for the third project, it’s required.
And I think that scaffolding really helps students develop that ethical muscle that I was talking about earlier: Just because you can use it doesn’t mean you should. And when they have that slower adoption and that guided adoption, I think responsible use increases at the end.
