Cooling Interest in Computer Science: AI Is Not the Only Reason

For a long time, computer science was seen as one of the safest career doors in the tech world. Good salaries, strong employment, big companies coming to campus and the chance to get a job right after graduation fed this perception. But recent data shows that this picture is now viewed more cautiously.

According to Harvard University's David J. Malan, explaining the decline in interest in computer science programs only through ChatGPT, Copilot or similar generative AI tools falls short. AI is undoubtedly raising new questions in students' minds, but there is a broader market story behind the decline: tech companies are no longer hiring as aggressively as they used to.

The Problem Isn't Just the Fear That "AI Writes Code"

Generative AI being able to write code has opened up the value of a computer science education to debate. Many students naturally ask: if a model can produce working code in a few seconds, does it still make sense to study computer science for four years?

The question is understandable, but on its own it isn't enough. The point Malan makes is this: the decline in student interest had started taking shape even before generative AI went mainstream. Tech companies that rapidly expanded headcount during the post-pandemic growth period have been more selective and more cost-focused over the last few years.

So students aren't only thinking "will AI take my job?" They are also facing the question "will I actually be able to find an entry-level job when I graduate?"

The Signal on Campus Has Changed

One of the most important factors that made computer science programs attractive was the visibility of big tech companies on university campuses. Career fairs, internship programs, graduate hiring and stories of high-paying first jobs were a powerful source of motivation for students.

Today that signal has weakened. Companies are opening fewer entry-level roles, some are participating in campus recruiting in a more limited way, and competition in the new-graduate market is getting tougher. This doesn't eliminate the long-term value of a computer science education; but it changes a student's short-term risk calculation.

In particular, the harder path into work for junior developers and new graduates is directly reflected in program choices. The sentence "study computer science and you'll get into a good company" used to be more convincing. Now the same sentence comes with more conditions: a strong foundation, real project experience, problem-solving skills, using AI tools correctly and being able to read the changes in the industry.

The Data Confirms the Decline

Data from the National Student Clearinghouse in the US also supports this cooling. According to data covering the vast majority of US universities, computer science enrollment at four-year colleges and universities fell by 8.1 percent in fall 2025.

This rate is striking not only for computer science but across academic disciplines in general. Being recorded as one of the biggest annual declines seen in any field since at least 2020 shows that this is not a passing rumor.

But reading this picture as "computer science is over" would also be wrong. The more accurate interpretation is this: computer science is no longer perceived as an automatic career guarantee. Students and families are putting the value of the major through a more realistic market filter.

Computer Science Is Bigger Than Memorizing a Programming Language

The most critical aspect of this debate is the perception of what computer science is. If the field is seen as nothing more than learning a particular programming language, AI tools really do look like a threat. Because models can now produce sample code, bug fixes and simple application scaffolding in many languages.

But that is not where the real value of computer science lies. Algorithmic thinking, abstraction, system design, data structures, security, scalability and problem-solving skills still require human judgment. AI doesn't make these skills unnecessary; on the contrary, it makes people who can spot badly generated code, understand the system and ask the right question more valuable.

That is why it is meaningful that figures such as OpenAI board chair Bret Taylor and Geoffrey Hinton emphasize the importance of computer science education. Knowing syntax alone won't be enough to succeed in the future, but that doesn't mean education is becoming unimportant.

A More Volatile but More Mature Period

Malan's prediction is that interest in computer science won't disappear entirely, but that we will enter a more volatile period. This is a quite reasonable expectation. The tech industry usually swings between excessive optimism and excessive pessimism. In one period everyone rushes to learn software, in another the idea that "AI will do it all anyway" spreads.

The balance will probably settle somewhere between these two extremes. Computer science still offers a strong foundation, but a diploma on its own is no longer enough. For students, the real advantage will come not from thinking better despite AI, but from thinking better together with AI.

So it is more accurate to read today's decline not as a sign of the end, but as a repricing of expectations. Computer science isn't losing its appeal; it is just moving from being a promise of an easy career to a more selective, more skill-focused and more realistic field.