UCSD - University of California - San Diego

04/07/2026 | Press release | Distributed by Public on 04/07/2026 17:29

What Does It Mean To Learn With AI

Published Date

April 07, 2026

Article Content

It's a familiar moment in many large college courses: a student working through a problem suddenly gets stuck. In Christopher Day's introductory genetics class at the University of California San Diego - which enrolls nearly 400 students - help from an instructor isn't always immediately available.

This year, they have another option. Day is piloting an artificial intelligence tutor created by researchers at UC San Diego and trained on materials from his course. Rather than providing answers outright, the system prompts students with questions and hints that nudge them toward a solution - giving them another place to turn while waiting for support from an instructional assistant or studying outside scheduled office hours.

"The goal isn't to replace the learning process," says Day, an assistant teaching professor in the School of Biological Sciences. "It's to help students work through the problem and move toward understanding."

Across higher education, universities are grappling with a central question - will generative AI undermine learning or reshape it? At UC San Diego, faculty members are exploring both sides of that question, developing AI tools designed to support teaching while preparing students to build the next generation of AI systems.

That work includes a new undergraduate major in artificial intelligence, launched in fall 2025. From the start, the program focuses on the mathematical and computational foundations behind modern AI systems while also examining the ethical and societal implications of the technology they are learning to build.

Together, the tutor and the new major reflect a broader direction at UC San Diego, championed by Chancellor Pradeep K. Khosla: treating AI not simply as a disruption to manage, but as something students and faculty can - and should - study, question and help shape.

Students at UC San Diego's Jacobs School of Engineering (Photos by Erik Jepsen, UC San Diego)

A tutor designed by instructors

The tutor used in Day's class is part of the SmartLearning Hub project, an initiative led by Mohan Paturi, professor of computer science and engineering and director of the Laboratory for Emerging Intelligenceat UC San Diego's Jacobs School of Engineering.

Paturi says the project began with a practical question shared by many faculty members about how to provide meaningful academic support to hundreds of students at once.

"The most important thing missing in large classes is individual attention," Paturi says.

The tutor is designed to extend that support. Unlike general-purpose AI tools like ChatGPT, it's built around course-specific materials and configured by instructors themselves. It follows a Socratic approach, prompting students to work through concepts rather than simply supplying answers.

"The instructor is really at the center of the system," Paturi explains. "It's built around the way the course is taught."

The tool is being used in courses across disciplines - including economics, data science, nanoengineering and biology - at several institutions in the San Diego region. Researchers are studying how students interact with the tutor and how it might fit into different learning environments.

In Day's genetics course, the tutor plays multiple roles. Students can use it while studying on their own, but they also have access to the tool during discussion sections, where it helps small groups work through challenging problems together instead of stalling while waiting for a teaching assistant to reach them.

Initially, some groups had access to the tutor for part of the quarter while others did not. With students' consent, some of the discussions were recorded so Day and his collaborators can analyze how students interact with one another - and with the AI - when they encounter difficult concepts.

"What we're really interested in is how students engage with it," Day says. "Do they use it to think through the problem? Do they ignore it? Do they challenge it?"

The results so far suggest that engagement varies widely. Some students utilize the tutor frequently. Others rarely open it. For students who do engage with it, Day sees value in having another source of support outside scheduled class time.

"In a class of this size, students can't always get immediate feedback," Day adds. "The tutor gives them another way to keep working instead of giving up or going straight to ChatGPT for the answer."

Christine Antoine, left, and Leena Banga, right, are first-year students in UC San Diego's inaugural AI major cohort.

Training future AI builders

While faculty members explore how AI might support teaching, the Jacobs School of Engineering is also expanding how students learn about the technology itself. The new undergraduate AI major, housed in the Department of Computer Science and Engineering, is among the first of its kind in the U.S. and was created in response to growing student interest and industry demand.

Mia Minnes, the department's vice chair for undergraduate education, says the program builds on years of investment in AI research both in the department and across campus.

"We saw enormous demand from students who wanted to go deeper into this area," Minnes explains. "At the same time, we were hearing from industry partners that specialized expertise in AI was becoming increasingly important."

The curriculum shares foundational coursework with computer science but introduces AI ideas early. Trevor Bonjour, an assistant teaching professor who helped shape the program and now teaches its first required course, says the curriculum follows a spiral learning model: students encounter key concepts and return to them repeatedly as their mathematical and technical skills deepen.

"We wanted students to be exposed to these ideas from the start," Bonjour says. "Then over time they revisit them again and again, each time with a deeper understanding."

Minnes says the program itself came together quickly due to both the demand and the depth of expertise already available on campus. UC San Diego intentionally launched the major with a smaller first cohort while continuing to build out upper-division courses.

According to Bonjour, the conversations happening in the classroom already reflect how quickly the field is evolving. Students regularly raise questions about bias in algorithms, the environmental cost of data centers and the societal impacts of AI technologies.

"They were already aware of a lot of these issues," Bonjour says. "In many cases the discussion became more about how they might address those problems in the systems they build."

A new generation entering the field

For first-year students like Leena Banga and Christine Antoine, both members of the inaugural AI major cohort, that sense of responsibility helped draw them to the program.

AI, Banga says, is already embedded in nearly every industry, from health care to transportation. That momentum helped convince her to pursue the field and be part of shaping what comes next.

"When you look at how many industries already rely on AI, it's clear this isn't something that's going away," says Banga. "And when you look at the people developing large language models today, there isn't a lot of diversity. Bringing in new perspectives helps ensure the systems we build are more responsible and less biased."

Her decision also points to a shift unfolding across higher education. For the first time since the early 2000s, enrollment in undergraduate computer science programs declined in 2025 across the University of California system. UC San Diego was the only campus to buck that trend, the same year it launched the new AI major.

Echoing Banga, Antoine adds that the program has also challenged common assumptions about AI itself.

"A lot of people think AI is easy because you can just type something into ChatGPT and get an answer," she says. "But the back end is really complicated. There's a lot of math and coding that goes into making those systems work."

AI beyond the classroom

For decades, UC San Diego has advanced AI research and education, including early work on neural networks that helped lay the foundation for modern systems such as ChatGPT. The campus has also developed TritonGPT, an in-house generative AI platform hosted on the high-performance computing infrastructure at the San Diego Supercomputer Center. Beyond the tutor and the new major, the campus is expanding conversations about artificial intelligence, including new AI literacy courses open to students across disciplines and work led by the Academic Integrity Officeexamining how generative AI is reshaping instruction and assessment.

Amid the nationwide debate over how AI should influence teaching and learning, UC San Diego's approach reflects a broader belief: that universities must do more than react to AI - they must help shape how the technology evolves.

UCSD - University of California - San Diego published this content on April 07, 2026, and is solely responsible for the information contained herein. Distributed via Public Technologies (PUBT), unedited and unaltered, on April 07, 2026 at 23:29 UTC. If you believe the information included in the content is inaccurate or outdated and requires editing or removal, please contact us at [email protected]