The great computer science exodus (and where students are going instead)

Something unusual played out across University of California campuses this fall. For the first time since the dot-com bust, computer science enrollment fell. Systemwide, enrollment dropped 6% last year after sliding 3% in 2024, according to reporting this past week by the San Francisco Chronicle. And this is happening even as overall college enrollment rose 2% nationally, based on January data from the National Student Clearinghouse Research Centre — a sign that students are stepping away from traditional CS degrees rather than college altogether.

There’s one clear outlier: UC San Diego, which was the only UC campus to introduce a standalone AI major this fall.

At first glance, the decline could look like a short-term wobble linked to headlines about fewer CS graduates landing jobs right after graduation. But it may be pointing to something more durable — and something China is leaning into with far more urgency. As MIT Technology Review reported last July, Chinese universities have aggressively prioritised AI literacy, treating AI not as a threat but as foundational infrastructure. Nearly 60% of Chinese students and faculty now use AI tools multiple times per day, and institutions such as Zhejiang University have made AI coursework mandatory. At the same time, top schools such as Tsinghua have established entirely new interdisciplinary AI colleges. In that environment, being fluent in AI isn’t a bonus anymore; it’s the baseline.

In the United States, universities are trying to catch up. Over the past two years, dozens of AI-focused programs have been rolled out. MIT’s “AI and decision-making” major is now the second-largest major on campus, according to the school. And as The New York Times reported in December, the University of South Florida enrolled more than 3,000 students in a new AI and cybersecurity college during its fall semester. The University at Buffalo, meanwhile, launched a new “AI and Society” department last summer, offering seven specialised undergraduate degree programs — and it drew more than 200 applicants before it officially opened.

But the shift hasn’t been seamless everywhere. When I spoke with UNC Chapel Hill Chancellor Lee Roberts in October, he described a spectrum of faculty reactions — some “leaning forward” with AI, others with “their heads in the sand.” Roberts, a former finance executive who came to the role from outside academia, was pushing hard for deeper AI integration even as some faculty resisted. Just a week earlier, UNC had announced plans to merge two schools to form an AI-centred entity, a move that drew faculty pushback. Roberts also appointed a vice provost dedicated to AI. “No one’s going to say to students after they graduate, ‘Do the best job you can, but if you use AI, you’ll be in trouble,’” Roberts told me. “Yet we have faculty members effectively saying that right now.”

Parents appear to be influencing the transition as well. David Reynaldo, who leads the admissions consultancy College Zoom, told the Chronicle that parents who once encouraged their kids to pursue CS are now nudging them toward majors they believe are less vulnerable to AI-driven automation, including mechanical engineering and electrical engineering.

Still, the enrollment trends suggest students are making their own choices. According to an October survey by the nonprofit Computing Research Association — whose members include computer science and computer engineering departments across a broad set of universities — 62% of respondents said their computing programs saw undergraduate enrollment decline this fall. Yet with AI programs expanding quickly, it’s starting to look less like a tech exodus and more like a migration. The University of Southern California is launching an AI degree this coming fall, and so are Columbia University, Pace University, and New Mexico State University, among many others. Students aren’t walking away from tech — they’re shifting toward AI-centred programs in hopes of improving their job prospects.

It’s still early to know whether this recalibration is permanent, a temporary panic, or simply a short-term response to a longer-term disruption. But it’s clearly a jolt for administrators who have spent years debating how to handle AI in the classroom. The argument over whether to ban ChatGPT is basically ancient history now. The real question is whether American universities can move fast enough — or whether they’ll keep arguing while students transfer to schools that already have a plan.