The University of California has agreed to pay $1.2 million to OpenAI to secure access to ChatGPT Edu for all its students and workers — a deal that has raised concerns among its faculty.
Read more California, Gov.-elect Becerra to face $2B budget shortfall, state analysts predict
While UC leaders say the purchase provides equitable access with built-in data protections at a low overall cost to the university, the Academic Senate has raised concerns over its exclusion from a decision that could have far-reaching implications for teaching and learning. This comes as instructors across the state call for greater caution in the use of AI in the classroom.
The yearlong agreement, which starts Dec. 1, includes access to the developer’s coding agent Codex, expanded workforce training and artificial intelligence literacy workshops, officials said.
Earlier, UC labs and campuses could purchase approved OpenAI services as needed under a pre-negotiated 2024 contract. That systemwide agreement did not provide services across the board to all campuses but only facilitated them. With this new expansion, the university is now offering the AI tool free of cost across the 10-campus system of 300,000 students and 265,000 workers.
“Our expanded partnership with OpenAI will ensure all members of the UC community have access to generative AI tools and will have their data protected if they choose to use them,” said Van Williams, chief information officer for the University of California Office of the President. “UC will continue to ensure AI resources are available to UC community members, while empowering faculty to decide which technologies they choose to support teaching and learning.”
Under the agreement, the UC will own the data saved to its licensed ChatGPT accounts and OpenAI will not be able to reshare, monetize or retain it, Williams said. That data will be subject to the Family Educational Rights and Privacy Act, a federal law that protects the privacy of student education records.
Faculty concerns
While UC leadership has promised “full discretion” for faculty over the use — or lack thereof — of AI tools in their courses, professors at the university say that is not enough. They should have been more substantially involved in the university’s OpenAI purchase decision, they say, because widespread AI access has implications for teaching, learning, assessment and research.
In a dated Oct. 30, Academic Senate Chair Susannah Scott raised concerns over “shared governance and the very limited Academic Senate consultation” in the decision to expand the agreement with OpenAI. Scott said the senate did not object to the use of AI or the idea of providing access to AI tools, and recognized the potential benefits of lower costs, improved security and more equitable access when purchased for the university as a whole.
However, faculty leaders who represent professors across the university had concerns and questions that were not engaged with or addressed prior to the purchase. For example, they feared that making a tool like ChatGPT available to all could be perceived as the university’s endorsement for its use in all courses.
“We do not view making ChatGPT Edu licenses available to every UC student, instructor, and staff member as a routine software procurement decision, but rather a significant academic policy matter that falls within the Senate’s delegated authority and expertise to advise on as part of our shared governance process,” Scott wrote.
At the , the UC’s AI Steering Committee revealed its plans for a systemwide AI strategy. That included contracting with AI developers such as Anthropic and OpenAI for coordinated access across campuses to their tools. At the time, the committee said any proposal would go through a privacy and security review as well as Academic Senate consultation before the university commits.
In her letter, Scott said Academic Senate leadership was informed about 10 days before the agreement was set to be finalized and only asked for formal input when they requested to contribute to the discussion. One professor, who represents the senate on the UC’s Chief Information Officer Council, had been included in prior discussions but that did not constitute real faculty participation, Scott said.
She asked that the purchase agreement be deferred to allow for meaningful consultation with faculty and for greater involvement in future decisions on buying AI tools for academic uses.
The Academic Council, which serves as the senate’s executive committee, put the OpenAI purchase agreement on the agenda for its Nov. 19 meeting. But by then, the decision had already been taken. In a letter to Scott Thursday, President Milliken said the university was moving forward with the purchase.
Read more Imagine biking from Folsom to Freeport. State, local leaders plan to link river trails
He said tools like ChatGPT were already being widely used by UC students and faculty through personal accounts not covered by UC-negotiated data protection terms. Milliken added that the provision of these tools “in no way mandates their use by students or faculty” and does not establish academic policy.
“After consideration of the points raised in your letter, I have decided that given the cost, access, and security and data protection benefits to UC community members, this purchase should move forward. This decision has the full support of all UC Chancellors,” Milliken wrote.
In a statement to The Sacramento Bee, Scott said she was surprised and disappointed by the decision which was announced only hours before faculty leaders were set to formalize their recommendations.
“Interactions between Academic Senate leaders and administrators over the past few weeks were brief informational exchanges, not substantive consultations. Unfortunately, the core concerns raised in our October 30 letter remain unresolved. Changes with such far-reaching consequences for research and instruction require direct engagement with faculty responsible for teaching students and directing their research projects,” she said.
Following in CSU’s footsteps
The faculty’s growing concerns over the entry of artificial intelligence into academics is not unique to the UC. Earlier this month, the California State University’s academic senate passed a resolution asking university leadership to focus its AI efforts on simplifying administrative tasks for faculty instead of encouraging its use for teaching and learning.
The CSU entered into a systemwide contract with OpenAI nearly two years ago in January 2025. At that time, the university paid the company $17 million for 18 months. This July, the university renewed that controversial contract at the cost of $39 million over three years. The renewal extended service beyond the university’s 470,000 current students and 63,000 workers at 22 campuses to graduates for the first year after graduation.
Since then, some faculty members have criticized the CSU’s financial investment in the contract at a time when campuses were laying off workers to balance their budgets. Students raised concerns over data privacy and environmental impact, and both groups of stakeholders asked for a seat at the decision-making table — a critique that the UC now faces.
The UC, in contrast to the CSU, took an overall slower approach to widespread AI adoption. That strategy may have aided in securing for the university the relatively lower $1.2 million price.
“The explicit starting point for us was that we’re not going to pay what the CSU paid,” UC chief information officer Williams said. “(We said) we have too much in the way of great research assets that you all (AI companies) have kind of benefited from to engage with this technology. There’s a huge value whenever a faculty member publishes a paper around gene discovery using your tools for it, and we should not pay what the CSU is paying.”
In its negotiations with OpenAI, the UC also “played hardball” and said it would go with Anthropic as its exclusive provider if OpenAI could not offer a better price, Williams said.
Does this mean the UC will not contract with Anthropic or other AI companies in the future? No, Williams said. In fact, talks with Anthropic are very much ongoing, and the UC has long had agreements with Google and Microsoft. The UC’s “multi-vendor strategy”, he said, aims to give students multiple choices with the “right level of protections.”
As the UC works towards subsequent contracts with AI companies, president Milliken promised in his letter to the academic senate that there would be “many opportunities” for professors and university leadership to discuss the use of the technology and how to lead the way in its “ethical and principled use.”
This story was originally published November 20, 2026 at 2:00 PM with the headline “UC inks $1.2M OpenAI deal to provide ChatGPT to all students; professors object.”
Read more Sacramento schools, restaurants and a statewide ballot measure. Here’s what to know
