Nearly Everyone’s Using AI at Cal State. And Nearly Everyone’s Worried About It.

Nearly everyone in the nation’s largest public university system is already using generative artificial intelligence. And almost no one fully trusts it.

That tension emerges from a sweeping new survey of more than 94,000 students, faculty, and staff across 22 California State University campuses, which finds AI deeply embedded in learning, teaching, and work — even as concerns about accuracy, ethics, job security, and long-term impact remain widespread. Cal State officials say it is the largest study of AI in higher education to date.

Ninety-five percent of respondents reported trying at least one AI tool, and most use it regularly at work, school, and in their personal lives. Students, 84 percent of whom use OpenAI’s ChatGPT, lead the way. Yet most respondents said AI-generated content must be verified by human research.

The survey results come amid debate over the Cal State system’s nearly $17-million deal with OpenAI to provide students and employees with access to a university-specific version of ChatGPT. That contract ends June 30. A petition with more than 3,300 signatures is urging Cal State’s chancellor, Mildred García, not to renew it. The deal was part of a broader push announced in February 2025 to increase AI integration across the system.

One key benefit of the new survey, said David Goldberg, an associate professor at San Diego State University and a lead researcher on the study, is that future decisions will be data-informed: “Knowing what tools people are using to make sure that when we purchase tools, we’re purchasing the right ones.”

The survey, distributed via email last fall, found that students are pragmatic but wary. More than half use AI regularly, and nearly two-thirds say it improves their learning. Yet 65 percent are skeptical about its role in education, and more than 80 percent worry about its implications for future job security. They are keenly aware of AI’s usefulness and its long-term trade-offs.

Students also report mixed messages in the classroom. Nearly 40 percent said professors encourage AI use, but only 33 percent said they were taught how to use it effectively. “One of the things we take from that is that it’s good to, at minimum, be clear with students about what expectations are, one way or the other,” Goldberg said.

Those faculty expectations are all over the map. In the survey, 7 percent report that they do not talk about AI, 22 percent discourage it, and 18 percent forbid its use entirely. Conversely, 6 percent require AI use and 19 percent encourage it; 28 percent say they are neutral. Two-thirds of teaching faculty include an AI statement in their syllabi.

More than half of faculty said they use AI to develop course materials, even as many expressed skepticism about AI’s role in students’ work.

Recent studies reinforce Cal State’s findings about mixed feelings regarding AI. A review of 31,000 course syllabi found that instructors are gradually warming to generative AI, increasingly allowing it for certain tasks like editing or study support while remaining restrictive in areas such as drafting original work. Yet another faculty survey indicates deep concerns about AI’s effects on students’ critical thinking, academic integrity, and the value of college degrees. Together, these reports suggest that universities nationwide are struggling to balance embracing AI as a tool with safeguarding learning outcomes.

At Cal State, a majority of staff, faculty, and students agreed that AI will become essential in most professions. Yet 82 percent of students, 78 percent of faculty, and 74 percent of staff said they are concerned about AI’s impact on job security.

One way that San Diego State has tried to combat such concerns, Goldberg said, is through its AI microcredential program, which provides comprehensive AI training to faculty, staff, and students and is now available across the Cal State system.

Among employees, staff in noninstructional roles showed the highest adoption and most positive attitudes toward AI, while faculty were the most conflicted: 56 percent see it as helpful, 52 percent as harmful. As a faculty member, Goldberg said he recognizes both sides, viewing AI as a productivity tool while worrying about the cognitive offloading that may come from overreliance.

But across all groups, there is strong demand for guidance. Eight in 10 staff members and seven in 10 faculty members want formal AI training; about half of students agree.

Cal State plans to repeat the survey next fall, Goldberg said.