US Computer Science Enrollment Drops 12.4% — Is CS Still Worth Studying?

Once a Safe Bet, Now a Gamble

Computer Science (CS) was long seen by countless students and parents as a "safe bet" — high pay, stability, strong demand. But according to data shared by a YouTube creator with a master's degree in computer science, that safe bet is being abandoned by a large number of students.

A national tracking dataset shows that this fall, computer science enrollment at four-year universities dropped 12.4%, following a decline last year as well. This isn't because students have stopped going to college: total enrollment barely changed. In fact, health majors at four-year colleges actually grew 6.2%.

For a high school senior, this shift sends a signal: what was once a steady choice now looks like a gamble. The creator admits that seeing these numbers made him sad — after all, he himself holds a master's degree in CS.

A Calm Reading of the Data: AI Isn't the Only Answer

It's worth emphasizing that these numbers are preliminary, and the tracker does not attribute the decline to AI. The creator explicitly points this out.

But that doesn't stop us from considering a more practical question: when AI can generate code, students and parents will naturally reassess the return on investment of a CS degree. The 6.2% growth in health majors also shows that students haven't stopped going to school — they're redistributing their major choices, moving from "code" to "care," from "certainty" to "another kind of certainty."

This shift itself is a signal worth watching. It reflects not only AI's impact on the job market, but also students' reordering of their sense of career security for the future.

It's worth adding that the source of the tracking data deserves attention. The "national tracking data" mentioned in the article typically comes from the Current Term Enrollment Estimates published each semester by the U.S. nonprofit National Student Clearinghouse Research Center (NSCRC). Its data is based on enrollment records from thousands of colleges and universities across the country — broad in coverage, but primarily preliminary estimates that may be revised mid- to late-semester. The report tallies enrollment changes by broad discipline category (CIP codes), which is why the comparison between Computer Science (CIP 11) and Health Professions (CIP 51) is meaningful. For this same reason, the report itself only presents trends and makes no causal inferences — it doesn't survey students' specific motives for switching majors, so "AI caused CS to cool" can only be treated as a reasonable hypothesis, not a data conclusion. Understanding this helps avoid misreading correlation as causation.

As for AI's actual impact on programming jobs, the public evidence remains complex. On one hand, tools like GitHub Copilot and Claude Code have indeed significantly improved the efficiency of code generation and completion, compressing entry-level, template-based programming tasks through automation; on the other hand, multiple industry surveys show that what companies lack most is engineers who can design systems, review AI output, and take responsibility for its quality. In other words, AI changes the structure of tasks rather than simply replacing jobs: repetitive coding decreases, while the share of architecture design, code review, requirements analysis, and human-AI collaboration rises. This structural change will take time to ripple through to high school students' major choices, which also aligns with the lag in enrollment data — students and parents tend to base decisions on media narratives and recent graduates' job prospects rather than real-time labor market data.

The Real Value of Studying CS: More Than Writing Code

One of the creator's most key insights is that his CS degree taught him far more than coding.

"My degree taught me to write code. It also taught me to think in systems, and that most problems have multiple solutions. Even if AI writes the code, that's the part I still want."

This is the deepest part of the whole discussion. Systems thinking, problem decomposition, multi-path problem solving — these abilities are actually scarcer in the AI era. AI can generate code, but it cannot replace the person who "knows what problem to solve" and "judges which solution is more reasonable."

The creator's regret is this: some students, in order to survive, have to give up these abilities and choose a major that looks "safer" right now.

"Systems thinking" is a concept often mentioned but rarely clearly defined. In the context of computer science, it means viewing a problem as a whole made up of interacting components, rather than an isolated line of code or a single function: understanding the interfaces between modules, how data flows through the system, where bottlenecks may appear, and how a local change can trigger global consequences. Paired with this is "decomposition" — breaking a vague, large problem into verifiable smaller ones — and "multi-solution solving," which means comparing the trade-offs of different algorithms or architectures under constraints such as time, cost, and maintainability. These abilities are scarcer in the AI era because generative AI excels at producing code once requirements are clearly specified, but struggles to judge for people whether the requirements themselves hold, whether boundary conditions are missing, and whether a solution can sustainably evolve in a real system.

The Paradox of Security: What Is Truly Safe

There's another intriguing detail in the data: total enrollment barely changed, while health majors grew 6.2%. This shows that students haven't fled higher education — they've fled "uncertainty."

But there's a paradox here: if everyone makes decisions based on what looks safe right now, today's "safe" may well become tomorrow's crowded track. The decline in CS enrollment may, to some extent, actually ease future oversupply.

The creator concludes by saying, "What looks safe right now is winning." This sentence is both a description of the status quo and tinged with worry — when security becomes the sole criterion for decision-making, students may sacrifice the abilities that truly have long-term value.

For students currently choosing a major, the key question may not be "Is CS still safe?" but "What kind of abilities do I want to gain?" If the answer is systems thinking and the ability to solve complex problems, then CS remains a choice worth serious consideration — regardless of whether AI is writing the code.

One-line Summary

CS enrollment plunges 12.4% as students turn to health majors, but the degree's value goes beyond writing code.

Paragraph Summary

Based on national tracking data shared by a YouTube creator with a CS master's degree, the article notes that computer science enrollment at four-year universities fell 12.4% this fall, following a decline last year, while total enrollment barely changed and health majors grew 6.2% — showing students haven't left higher education but are redistributing their major choices. The author stresses that the data is preliminary and does not attribute the decline to AI, but AI's ability to generate code is indeed prompting students and parents to reassess the ROI of a CS degree. The article's core insight is that the real value of a CS degree lies in systems thinking, problem decomposition, and multi-solution problem solving — abilities that are scarcer in the AI era. The author warns of the "security paradox": when everyone chases today's safe option, today's safe harbor may become tomorrow's crowded track.