Schools across the University of Texas at Austin are integrating a new AI tutor, “UT Sage,” to enhance learning.
AUSTIN, Texas — As artificial intelligence keeps changing, more teachers are starting to integrate it into their classrooms. This year, several schools across the University of Texas at Austin campus will be using a new AI tutor platform called UT Sage.
Nina Telang, a professor in the Cockrell School of Engineering, has been teaching at UT for more than 20 years. But because of new AI technology over the last few years, her courses have evolved.
“For me to be a better teacher, I need to immerse myself in their world and kind of embrace it, just like they have embraced it,” Telang said.
Over the last year, she’s been using an AI platform created by one of her former students that’s similar to the new campuswide one, Sage. The AI tutor is trained on her course materials through material she uploads, like her notes, lecture slides, homework assignments and problem solutions.
“The student can ask a question like, ‘Can you please explain to me this topic? Break it down for me in steps,'” Telang said. “The idea that this AI tool can be used as a tutor, that is, I think, a huge value for the student.”
She said the tool has been a great help to supplement what students learn in the classroom.
“Think of like, a student who is feeling shy to ask a question in the class. Or in the classroom, you know, I need to cover x, y and z concepts in the class because I have 75 minutes and then we move on,” Telang said.
Maintaining academic integrity with new AI platform
With the new innovation comes some challenges. Telang said she’s had to rethink the way her course is designed a little bit.
In previous years, Telang said her courses were split evenly between assignments and assessments. Now because of all the AI assistance, it’s more of a 30-70 split, with more emphasis on testing what they’ve learned from the platform.
“If you cheat in that 30, 40%, then you’re not going to be able to really do well in that 60-70%,” Telang said. “I need to make sure that these are the skills that they have at the end of the semester.”
Her hope is that the AI assistance will give students the help they need to learn the material outside the classroom, while the instructional time is where they build other skills.
“This is how you build that foundation, and that knowledge base, and those critical thinking skills,” Telang said. “It comes from classroom discussions, it comes from their peer-to-peer interactions.”
UT’s AI principles
With more AI innovation always emerging, UT made updates to its AI framework this spring. According to the university, they have eight principles for using AI responsibly for teaching and learning.
“We want students to innovate and use and adopt technology, but we want them to do it in a way that is ethical and has integrity,” said Julie Schell, the assistant vice provost of academic technology.
Below are the principles, as outlined by the Office of the Executive Vice President and Provost:
- Literacy: Use AI with fluency in its principles, applications, limitations, and impacts in higher education and beyond.
- Intention: Use AI transparently and in the support the achievement of learning in and across disciplinary contexts, including knowing when and why it should and should not be used.
- Balance: Use AI with an innovation and future readiness mindset, and with the awareness of the evolving benefits and drawbacks in the context of teaching and learning across areas of study.
- Agency: Use AI in ways that maintain and enhance the agency of humans over their intellectual output, decision-making, and the teaching and learning process.
- Ethics: Use AI in ways that recognize the array of ethical implications of technology use on the environment, labor, society and other considerations, engage with those implications and consequences across contexts, and consistently practice evaluating when AI can versus should be used.
- Relationships: Use AI in ways that enhance and extend the connections between everyone within a learning community—students and their peers, faculty and students, staff and the people that they serve—rather than diminish or replace them.
- Academic Integrity: Use AI in alignment with our honor code and fundamental scholarly values such honesty, respect, and authenticity.
- Stewardship: Use AI in ways that are explicit about how data is being shared, and guardrails are established to uphold the privacy, safety, security, accessibility, intellectual property rights and right to access of everyone in our teaching and learning community.
Among those principles, Schell said one that was brought up many times between faculty and students was relationships.
“Using AI in ways that foster those connections, rather than diminishing or replacing them,” Schell said. “So, I think one of the interesting things that we’ll do is really trying to lean into the importance of that relationship.”
The university also has developed what it calls the “Big 6“, which are the six key limitations of using AI for learning. Those are privacy and security, hallucinations, misalignment, bias, ethics, and cognitive offloading.
UT’s working group for the AI framework accepted public comment on the principles this spring through July 31, but Schell said people are always welcome to keep submitting feedback to them, as technology continues to change.