Faculty Symposium Highlights AI’s Strengths in Higher Ed Teaching

Story by:

• Catherine Poland - [email protected]

Share This:

• Share this story on Linkedin
• Share this story on Facebook
• Share this story on Threads
• Share this story on Twitter
• Share this story via email

Article Content

Intellectually curious. Engaged. Creative. Collaborative.

Leaders.

This is how Chancellor Pradeep K. Khosla described the University of California San Diego faculty who presented at the faculty artificial intelligence symposium in late May, sharing the cutting-edge solutions they have built to integrate AI tools into their teaching and expand student learning rather than limit it.

"UC San Diego was created as a campus to experiment with pedagogy and education and what the future would look like. So, there's nothing better that this campus can do than look at AI," said Khosla as he opened the symposium. "I want us to lead this conversation."

As co-chair of the recently announced AI steering committee for the University of California, Khosla is guiding the system's exploration of the ethical, educational and societal issues AI poses for its 10 campuses, the state, the nation and beyond.

"We cannot stop the tsunami. The genie is out of the box," Khosla said. "The best thing for an educational institution to do is to understand how we as educators can understand the effects of AI and prepare the current generation and the next generation to live and work in this world."

How is UC San Diego leading the way?

There is a growing global argument that the overuse of AI can soften critical thinking skills while simultaneously offering people an illusion of confidence, particularly when they accept AI generated content without questioning it. At the symposium, faculty showcased methods and tools each had developed to counter this erosion of problem-solving skills in students.

A throughline in the day's presentations also included practical solutions to train students to use AI thoughtfully so that they develop, retain and grow their thinking, knowledge and skills as part of their degrees.

The tools and strategies ran the gamut from newly developed AI tutors, randomized question banks and more, to using AI as a way to interrogate faculty themselves, provide feedback on student learning, and help organize thoughts for lecture writing.

When it comes to AI tutors, Sesh Murthy, a strategic research initiatives specialist from the Halıcıoğlu Data Science Institute in the Halıcıoğlu School of Data Science and Computing, and a team have created a novel solution that has significantly improved pass rates in UC San Diego's Foundations of Precalculus course.

Murthy pointed out that a study from the Khan Academy has shown that typically only 5% of students use tutoring enough to benefit from it - and these are the students who probably need help the least. To better reach students who need the help that a tutor provides, the team developed the AI-based Adaptive Scaffolding for Personalized Instruction and Responsive Education, or ASPIRE. This approach embeds AI learning support into crucial course tasks such as homework, practice and test preparation assignments rather than relying on students to seek out optional help on their own.

With testing across 12 courses including more than 2,200 students, the ASPIRE team found that outcomes were improved with the comprehensive interventions that included AI. Benefits included the fact that AI support was available 24/7 and seen by students as offering a safe space to ask basic questions. Additionally, the AI tutors caught mistakes in real time and generated instant, personalized feedback and individualized foundational review units for students.

Crucially, these AI tutors were designed to preserve students' "productive struggle" as they learned skills through practice and without simply being given the answers, which aided in their retention of the content, according to Murthy.

Preliminary results are positive, Murthy pointed out. The failure rate in Foundations of Precalculus dropped from 35.3% in Fall 2024 before ASPIRE was integrated into the instruction to 11.4% in Fall 2025, after a comprehensive instructional redesign that included the AI tutor.

AI can also be used to support assignment-based learning, said Vincent Nijs, a teaching professor in the Rady School of Management. In his courses, Nijs is developing AI coaching tools that guide students through required conversations as they complete assignments. The system tracks student progress, identifies gaps or misconceptions, and offers targeted prompts rather than simply providing answers. This creates an apprenticeship-style learning model at scale, giving students timely feedback while helping instructors focus class time and individual conversations where they are most needed.

AI tools are also useful when it comes to replicating practice questions to deepen learning. Teaching Professor Mia Minnes, in the Department of Computer Science & Engineering, wanted to generate a robust question bank to use with PrairieLearn, a homework and practice software tool that UC San Diego licenses for faculty to use for instruction. Minnes, who also serves at the department's vice chair for undergraduate education, found that AI was a valuable tool to easily change what she calls the "window dressing" of questions so that - through repetition and practice using her randomized question bank - students have the opportunity to review and strengthen core concepts they need to practice with rather than "mistaking memorization for learning," as she explained it.

Benjamin Noble, an assistant professor in the department of political science in the School of Social Sciences, explained how he used AI to invert the typical process for writing lectures. He has developed a workflow that interviews him about whatever topic he identifies for a lecture, identifies gaps in planned topics and produces a structured outline for the lecture. Rather than a shortcut, he noted that his method was effectively a "long cut" because of the time it takes, but the process deepens the content of his lectures and ensures uniformity of structure while maintaining a clear focus on a course's designated learning outcomes.

What are the next steps at UC San Diego?

At the end of the symposium, Khosla unveiled two new initiatives for faculty interested in integrating AI tools into their teaching. The first is the Citizen Developer Initiative, which is a bank - much like an app store - that UC San Diego Information Technology systems is developing so that faculty members can upload what they have developed for others to utilize, build upon and securely test for their own use.

This centralized Triton AI initiative "puts more powerful AI exploration directly in your hands," Khosla shared. "I think what's missing between now and great success is going to be scale, and each faculty member gets to define what the scale is for themselves: How many people implement or adopt what you are doing? How many students are being impacted?"

The centralized bank enables faculty to build upon the successes of their colleagues, so that personal productivity tools have the potential to become campus-wide solutions.

The second initiative is an AI Teaching Commons, which is being developed in collaboration with faculty and staff, Khosla shared. "This will help build AI literacy at scale across our campus community and ensure that students, faculty and staff have the support needed to engage these tools effectively and confidently," he noted.

The AI Teaching Commons partners together the Teaching and Learning Commons, Academic Technology Services, the Division of Extended Studies, and Operational Strategic Initiatives to support community members as they learn to use AI tools effectively and confidently in their workflows.