Cal State has 58 days to decide a $16.92M OpenAI deal. The numbers on the ground say don’t.

By Sky | type0

Eighteen months after California State University signed a deal with OpenAI, only 0.7 percent of its students and 16 percent of its faculty have completed the voluntary training the university built for ChatGPT Edu. The system's own survey of more than 94,000 students and employees found 52 percent of faculty reporting that AI had a negative effect on their teaching. And in 58 days, Cal State will have to decide whether to renew — without having said publicly whether it plans to.

The system's survey found 67 percent of students saying their professors do not teach them how to use AI effectively — even as 64 percent said AI had affected their learning experience positively, and 63 percent had seen more opportunities to learn about AI on their campuses (CalMatters; SDSU AI Survey Dashboard).

"We were not consulted when the contract was signed, and we weren't even given a heads up," said Katie Karroum, a communications major at Cal State Northridge and vice president of systemwide affairs for the Cal State Student Association (CalMatters).

Cal State signed the $16.92 million agreement with OpenAI in January 2025, giving all 23 campuses free access to ChatGPT Edu through single sign-on by April. The contract was signed January 17 by Chief Procurement Officer David Beaver on behalf of the Board of Trustees. Cal State chose OpenAI as the least-costly option among vendors it evaluated, according to assistant vice chancellor of academic technology services Leslie Kennedy — but left training voluntary (The Orion; Cal State Press Release; CalMatters).

Some faculty pushed back hard. Ryan Jenkins, a philosophy professor at Cal Poly San Luis Obispo and chair of the campus faculty union's AI task force, has returned to in-class bluebook exams using scantrons specifically to avoid ChatGPT during assessments. When ChatGPT first came out in 2022, he said, he tested it by giving it a reading quiz. It answered every question correctly. "The bread and butter of philosophy is reflecting on your own ideas and trying to sort out what you believe and why," Jenkins said. "If you have a tool that does that for you, then you're being denied an opportunity to practice that skill" (CalMatters).

More than 3,000 Cal State students, faculty and alumni had signed a petition as of April calling on Chancellor Mildred García not to renew, directing any savings toward protecting jobs facing layoffs (Action Network). SFSU professor Martha Kenney put it bluntly: "Introducing generative AI, which is not an educational technology, into a university system that is really, really crumbling under austerity right now is just a recipe for disaster. To put a broken technology into a broken system can only break it further" (Tech Policy Press).

The opposition sharpened when OpenAI announced a deal with the Department of Defense in March. Organizers said 250 people signed the petition in the two days that followed (Tech Policy Press). Assemblymember Mike Fong introduced Assembly Bill 2392 in February, which would require Cal State and California Community Colleges to provide structured training on any AI product deployed on their campuses — a direct legislative response to the rollout's ad hoc character (CA Assembly Bill 2392; CalMatters). At a joint legislative hearing last August, representatives from the Academic Senate, Cal State Student Association, California Faculty Association and Cal State Employees Union testified to their discontent with how the deal was made and what it had produced.

Cal State's chief information officer Ed Clark, who oversaw the procurement, defended the approach at that hearing. "We can't just sit back and let it go by," he said. "The best way to deal with those concerns is to have our universities participate in helping to shape the future of these technologies" (CalMatters).

Clark said the system had hired a firm to assess its return on the OpenAI investment, evaluating factors including classroom usage, student learning outcomes, administrative productivity and environmental impact. Cal State has not announced whether it will renew the contract, and a spokesperson did not respond to questions about the status of the renewal decision (CalMatters).

One complication: the 0.7 percent training completion figure measures only AI Commons, the voluntary portal the system built. It does not necessarily reflect how many people accessed ChatGPT Edu directly — Cal State students and faculty have had automatic access to the platform since April 2025 through single sign-on, regardless of whether they completed any AI Commons modules. Actual usage figures, if they exist, have not been made public.

The AI Workforce Acceleration Board that Cal State formed alongside the deal includes representatives from OpenAI, Adobe, Amazon Web Services, Anthropic, Google, IBM, Intel, LinkedIn, Microsoft, Nvidia and Soundings — the same companies positioned to benefit from whatever AI skills Cal State students develop (The Orion). The board is meant to identify "AI skills needed in California's workforce," according to its website.

The June 30 expiration is now 58 days away. What Cal State decides — and what it says about why — will be the first real data point on whether large-scale institutional AI procurement can recover from a rollout this rocky, or whether the gap between purchase and adoption is structural.