The Great Switch – Why Undergraduates Are Fleeing Traditional Majors Because of AI.

It feels different to stroll through Stanford’s computer science building on a Tuesday afternoon. Compared to three years ago, the hallways are now quieter. There’s a different vibe to the lines outside advisors’ offices; there’s more anxiety, less excitement, and more people holding printouts of course catalogs from departments they never would have thought to look into.

There is a change taking place on American college campuses, and it is happening more quickly than most institutions are ready to deal with. AI has already caused one in six college students to switch majors. Almost half have given it some serious thought. When you sit with the numbers, they are startling.

For nearly twenty years, computer science was the degree that parents boasted about at dinner parties and that high school counselors strongly advised. It was useful. It was profitable. According to the argument, it was recession-proof.

As quickly as construction budgets permitted, universities expanded their computer science programs, and enrollment grew year after year, with students arriving with laptops and partially completed LinkedIn profiles. For fifteen years in a row, the computer science department at the University of Minnesota expanded. The typical cycles of academic fashion didn’t seem to affect Stanford’s program. Then, all of a sudden, it wasn’t.

According to the Computing Research Association’s October report, enrollment decreased in 62% of traditional computer science programs this fall. The irony at the heart of this is almost too clever: the career path that those degrees were supposed to open up has started to be undermined by the technology that students were supposedly being trained to build.

After over 150,000 layoffs in 2024, over 100,000 tech workers lost their jobs in 2025. These losses weren’t limited to startups. Thousands of jobs were eliminated by Amazon, Meta, Google, and Microsoft, many of which were held by engineers in their early careers. Pupils took notice. They consistently do so more quickly than institutions acknowledge.

In a way, what’s taking the place of CS on the enrollment charts is a very sensible reaction to an absurd circumstance. In its first semester, more than 3,000 students were enrolled in the University of South Florida’s new College of Artificial Intelligence and Cybersecurity. With almost 330 students enrolled, MIT’s “Artificial Intelligence and Decision-Making” program is currently the second-largest major at the university.

Between 2020 and 2024, the AI master’s program at SUNY Buffalo expanded twentyfold. These are comprehensive programs with significant student demand that were developed almost entirely in response to a labor market that abruptly changed under everyone’s feet. They are not small experiments tucked away in elective catalogs. Observing all of this, there’s a sense that higher education is doing something unique: rather than waiting ten years to update a curriculum, it is actually reacting to the market in something that is almost real time.

These choices are motivated by specific, tangible anxiety rather than nebulous technophobia. According to data from SignalFire, the percentage of graduates from prestigious engineering schools, such as MIT, Stanford, Carnegie Mellon, and Berkeley, working as engineers at large tech companies fell from 25% in 2022 to just 11–12% today, a decrease of more than 50% in just two years. In group chats and dorms, that figure circulates like a cold fact that no one wants to dispute.

Students who apply for internships talk about competing with laid-off mid-level developers who suddenly need entry-level work, submitting hundreds of applications, and waiting six months for a response. Students are aware that the entry point has narrowed to the point where a degree is insufficient. For many technical positions, major companies like Google, IBM, and Accenture have eliminated the need for a degree, favoring proven abilities over credentials. In other words, the credential is no longer guaranteed.

All of this could be interpreted as simple panic, with students chasing the next big thing in the same way they chased degrees related to cryptocurrencies a few years ago or switched to finance after 2008. That skepticism isn’t totally unjust. The same displacement pressures that are currently reshaping computer science are not guaranteed to be avoided by anyone majoring in AI.

The warning that “rapid skills change and knowledge turnover may mean formal degrees are more rapidly out of date” in PwC’s 2025 Global AI Jobs Barometer report applies to AI degrees just as easily as it does to other degrees. By the time they graduate, students who rush into AI programs might discover that the field has changed once more, that the tools they studied are out of date, and that employers are once again looking for something different.

And yet. By admitting that the previous map is incorrect and attempting to create a new one under duress and with insufficient information, students are acting in a way that is genuinely logical, perhaps even courageous.

A generation of tech commentators made the confident “learn to code” promise, but it has backfired in certain documented ways. 3.2% of philosophy majors, only 3% of art history graduates, and 4.4% of journalism majors are unemployed; five years ago, these figures would have seemed absurd in comparison to CS projections, but they no longer do. Campus recalibration is underway. Whether the new course is the right one is still up for debate. However, it’s difficult to ignore the fact that students who are posing challenging questions about their futures are doing so with a clarity that many of their instructors, who continue to teach the same curricula, have yet to match.