For years, informatics was the safest choice at university. You learned to code, and you got a good job. That deal is breaking. In the US, undergraduate enrollment in computer and informatics sciences fell 8.4% in spring 2026. Students are not leaving tech, though. They are leaving the old way of teaching it.
I think universities should stop fighting this. Traditional informatics courses, built around writing code line by line, should end. They should be replaced by programs where AI is the starting point, not an extra subject at the end.
Students already decided
The data shows where students are going in. In Connecticut, the biggest drops were in traditional courses like coding and rebotics, while interdisciplinary computer science courses actually grew. In Isreal, computer science fell 7.4%, but AI, data science, and data engineering went up by 3.9%. At MIT, the new “AI and decision-making” major is now the second largest major on campus. Even at Kozminski University, the program where I’m currently studying is becoming more popular year by year: “Management in AI.” Students aren’t lazy. They are reading the job market better than the universities.
China is moving even faster. Zhejiang University made AI coursework mandatory, and Tsinghua created new interdisciplinary AI colleges. If European universities keep teaching informatics like it is 2015, our graduates will fall behind.
What companies actually need
This is a management issue, not only an education issue. Companies do not need juniors who type code that AI writes in seconds. They need people who can manage the AI itself: give it the right tasks, check its output and connect it to the business’s goals. IMB already understood this. It plans to triple entry level hiring in 2026, but it rewrote job descriptions to focus less on coding and more on working with customers. If companies redesign jobs this way, universities have to redesign courses too. Otherwise they produce graduates for jobs that no longer exist.
Don’t throw everything away
I have to be critical of my own argument here. Some experts warn that programming skills are still essential, especially to understand what AI can’t do. I agree. You can’t check AI-written code if you have never written any code yourself. So “ending” old informatics does not mean deleting the basics. It means changing the order. Students should learn fundamentals through AI tools from day one, instead of spending two years on manual coding before they even touch AI. My own course works like this: we learn the basics of coding, but not the old way. The old way is less efficient and is not what big tech companies are looking for. They want someone who knows what they’re doing and produces results. If you combine AI with coding knowledge, you get someone who is ten times more efficient than a normal programmer.
My Opinion
Fortune presents the drop in enrollment as a crisis for computer science. I don’t agree. I see it as a correction. For years, students picked informatics just because it promised a safe, well-paid job. Now that AI changed the rules, only people who are really interested will stay, and they will push the field forward. As a student in the management in AI program, I see this myself. The most interesting work is not writing code but deciding what AI should do and checking if it actually works.
Conclusion
Universities are organizations, and like any company, they must adapt or lose their customers. Students already use AI every chance they get. Informatics will not die, but it has to change. The schools that put AI at the center of their courses will produce the graduates that future tech companies will hire. The ones that don’t will keep teaching skills that are no longer in demand
References:
- https://fortune.com/2026/08/06/computer-science-enrollment-plunging-ai-college-campus/
- https://www.govtech.com/education/k-12/is-ai-leading-to-declining-enrollment-in-computer-science
- https://www.timesofisrael.com/amid-ai-shifts-and-war-israel-sees-drop-in-math-and-computer-science-student-enrollment/
- https://pastmeetspixel.substack.com/p/the-changing-value-of-coding
- https://techcrunch.com/?p=3092703
AI usage: Claude Opus 5.5
Prompts:
- “Give me some topics for my tech blog.”
- “Is making informatics more AI-focused a good topic?”
- “Can you check if my points are well expressed, in the right order, and connect nicely?”
- “Fix all grammar issues and list the reference links I provided.”