Princeton U. Gives Campus ChatGPT Access As Educators Warn Of AI Overreliance

PRINCETON, NJ — Princeton University students will be able to use ChatGPT through a university-managed account starting in January, but first they’ll need to complete a training module on AI use.

The university announced the Princeton ChatGPT pilot on Wednesday. Faculty, academic professionals and other staff can request access starting now. Graduate and undergraduate students can request access in January, before the spring semester begins.

The move puts Princeton among a growing number of universities offering AI tools campus-wide, even as faculty across the country raise concerns about what the technology is doing to how students learn.

The pilot comes from an enterprise agreement Princeton negotiated with OpenAI. It lets the university offer ChatGPT Plus and Pro accounts to faculty and staff “at a fraction of market rates,” Provost Jennifer Rexford said in a message to faculty and staff. It will also give researchers access to OpenAI’s most advanced models, she said.

Under the agreement, OpenAI cannot use content submitted by university users to train its models, Rexford said. The arrangement also offers stronger data security and privacy protections than individual accounts, and lowers costs for departments and researchers already using OpenAI’s tools, she said.

The university will evaluate the pilot after one year to decide whether to keep offering campus-wide licenses.

The university scheduled the student rollout for January to give faculty time to think about how campus-wide access to ChatGPT may affect their teaching, Rexford said.

Even with an account, students won’t have free rein. Faculty, departments and university policy will still decide when and how students can use AI in their coursework. Restricted and highly restricted information should never be entered into ChatGPT or other generative AI tools, Rexford said.

“I know the topic of AI at Princeton and in higher education is complex and raises many questions, and I am grateful to the faculty and staff who have provided input as we developed this pilot initiative,” she said.

The pilot is a step up from Princeton’s earlier approach. When the university rolled out Microsoft Copilot in May 2024, it was the only generative AI tool the Office of Information Technology offered. In January 2023, shortly after ChatGPT’s release, university deans told faculty Princeton did not plan to ban it.

Other Schools, Other Approaches

The California State University system, the nation’s largest public four-year system, offers ChatGPT Edu to more than 460,000 students and 63,000 faculty and staff across its campuses. The original 18-month deal cost nearly $17 million, LAist reported, and the system later renewed it for three years at $13 million a year, according to EdSource.

The contract has drawn pushback. Faculty petitioned to cancel it, and students protested on CSU campuses this spring, citing cost, environmental concerns and the company’s priorities.

Training has been a sticking point there too. Only 0.7 percent of CSU students and 16 percent of faculty completed the system’s voluntary AI training, CalMatters reported. Princeton is making training a requirement before students can get an account.

Duke University launched a pilot giving all undergraduates free, secure access to ChatGPT-4o.

Educators Have Doubts

A national survey of 1,057 faculty members, conducted in November 2025 by the American Association of Colleges and Universities and Elon University, found widespread worry. Ninety-five percent said generative AI will increase students’ overreliance on the tools, and 90 percent said it will diminish critical thinking.

Seventy-eight percent said cheating had increased on their campus since the tools became widely available, and 68 percent said their colleges had not prepared faculty to use AI in teaching, according to the survey. The survey was not scientific.

The debate isn’t limited to college. A Brookings Institution report released in January concluded that the risks of generative AI in K-12 education currently outweigh the benefits, NPR reported. The report cited evidence that students using AI are seeing declines in content knowledge, critical thinking and creativity, though it also found that teachers who use AI save nearly six hours a week.

In Washington, Sens. Richard Blumenthal, D-Conn., and Josh Hawley, R-Mo., wrote to Google CEO Sundar Pichai on Oct. 21 with questions about how the company’s AI products affect children in classrooms. They asked for answers by Nov. 2.

Princeton University did not say how much the pilot costs, whether students or staff will pay anything, or how many licenses are available.

Have a correction or news tip? Email sarah.salvadore@patch.com