Over the past three years, generative AI has gone from an illicit homework helper used by students to a technology embraced by higher education institutions. Last year, the California State University system entered multimillion-dollar deals with multiple companies including OpenAI, Amazon and Nvidia to embed their AI technologies into learning and student development. 

The California Community Colleges system also collaborates with Google to supply its 2.1 million students and faculty with AI tools. Additionally, at least half of UC campuses have developed customized models, such as UC Irvine’s ZotGPT. 

Recognizing the ways in which AI is already reshaping education, UC Santa Cruz established an AI Council at the beginning of 2025. The council comprises faculty, staff and one graduate student, all from a variety of different academic backgrounds, with the purpose of “guiding the university’s approach to AI integration,” according to the council’s mission statement.

Yet with little money to invest in AI initiatives due to the $79 million budget deficit, UCSC has more time to deliberate its intentions.

“I think the cost would be significant,” said former Vice Chancellor of Information Technology Aisha Jackson, who left her position last quarter to join UC Davis as their Chief Information and Digital Strategy Executive. “We’re not being as bullish in this space as other campuses might be, given the financial state of the campus, quite frankly,” she said of UCSC’s process. “We’re trying to be really studied and measured in our approach.”

Within this exploratory period, there are decisions that could challenge the ideals that UCSC prides itself on, and change the future of education for its students. 

Current AI Integration Efforts

UCSC’s Information Technology Services (ITS) added a chatbot to the Financial Aid and Scholarships Office (FASO) website in fall 2025 to assist students with basic questions. 

“We’re really looking hard at the options that we have internally or through existing licenses,” said Assistant Vice Provost for Educational Innovation Michael Tassio, who co-chairs the council. 

Within this absence of initiatives, some professors have experimented by creating their own tools. 

Adam Smith, an associate professor of computational media and a member of the council, has created an AI model called Brace that integrates with Canvas and enables students to complete assignments by working with the AI. Because Brace is trained on course learning materials and Smith stores the chat logs only on his system, he finds it more useful and private than using an already existing external model.

Leilani Gilpin, another professor within the computer science and engineering department, is also developing an AI autograder that can explain its decisions. Computer science students may already be familiar with using AI autograders to submit their work, and Gilpin’s design would add insight into the AI’s grading process, making it more valuable for students and teachers.

“One of the attractive things about AI graders — at least from an instructor’s perspective, not my own perspective — is that it would give time to the TAs and the instructors to actually teach the material instead of just grading, which I think is a common issue,” Gilpin said.

According to Smith, when his class size was increased and sections canceled because of room scheduling conflicts, his existing plan to implement Brace “helped us scale the course in ways we couldn’t before.”

The majority of Smith’s students used Brace for assignments, but some students like Ashley Seward, a fourth-year computer science game design major, dislike the use of Smith’s model in the class, and feel that it’s not effective to their learning. 

“I don’t really think that it’s helpful that we have a semi-functional system that evaluates our work,” Seward said, describing how Brace would sometimes give faulty feedback on her assignments. “I would rather have another TA, but that’s not in the hands of Professor Smith.”

While Aisha Jackson and Michael Tassio stressed that the AI tools UCSC has enabled can help staff and administration do their jobs more efficiently, the question of downsizing staff arises. Students have already begun to see cuts across departments, from college advising office mergers to FASO staff terminations. 

Problems with AI

Adam Smith hopes that UCSC can eventually adopt a model similar to what he uses now, instead of buying into expensive deals with outside vendors whose models might be less adapted to classroom needs, and even a privacy concern. 

College students, according to Smith, represent an ideal customer base for companies like OpenAI, as their models are trained using the data fed to it. The ways college students use AI represent an attractive new market to companies hungry for usage data.

“They want to know what their customers are doing,” Smith said. “If you can get people hooked on your technology when they’re young, they may be dependent on it later. Our data is very valuable to them.”

There’s also the fact that AI data centers pollute the regions they are located in and use large amounts of energy and fossil fuels. UCSC admissions uses the nature on campus to appeal to prospective students, and the school often advertises itself as environmentally-conscious. What happens to this image and ideal when, as UCSC is exploring, AI may be integrated into classrooms across campus? 

“The main places that AI is affecting in terms of the impact is the global south,” Gilpin said. “We’ve said that we care about green energy and climate and all those things, so that should absolutely be a part of the conversation.”

What’s Next?

Some professors like Smith and Gilpin feel strongly about making them open-weight, which allows users to download, run and customize AI models. They also want to ground experimentation in education work. Many students are concerned about how AI could harm their education and careers. 

“We need all instructors and students to be an active player in these conversations,” Michael Tassio said. “Being co-chairs of the AI council, we do hear from undergraduates through the form that we have on our website and we get messages like, ‘Ban AI. There’s no place for AI in school.’ We hear all sides of it.”

In October 2025, the council sent out a campus-wide survey on AI that will be collated into a publicly available report at a later date. ITS the hiring of a service manager for AI, of which the search is currently underway, and the university is also seeking an undergraduate representative to join the council. 

The next priority of the AI council is mapping out a call for proposals for AI pilots and considering which divisions to implement them in.

As other schools may barrel forward with this technology, it’s clear UCSC stands at a crossroads, and is on the precipice of making big decisions about AI. 

“We are going to exist in a future where everyone around us is using these tools,” Tassio said. “What I’m most worried about is assuring that the people who are using all of these tools are still able to think really creatively and critically and analytically about how they’re using them.”