As artificial intelligence (AI) systems expand their already impressive capabilities, it’s becoming increasingly common to think that the field of computer science will soon be a thing of the past. This is passed on to prospective students in the form of well-intentioned advice, but it’s largely hearsay from people who, despite their intelligence, speak outside their field of expertise.

Renowned figures such as Nobel Prize-winning economist Christopher Pissarides have made this argument, and as a result it has taken root on a much more mundane level: I myself have heard high school careers advisors dismiss the idea of studying computer science, despite having no knowledge of the field itself.

These claims often share two common flaws. First, the advice comes from non-computer scientists. Second, there is a widespread misunderstanding about what computer science actually entails.

AI and the myth of code replacement

It’s not inaccurate to say that AI can write computer code from prompts, just as it can generate poems , recipes, and cover letters. It can increase productivity and speed up workflow, but none of this negates the value of human input.

Writing code is not synonymous with computer science. You can learn to write code without having attended a single college class, but a degree in computer science goes far beyond this skill. It involves, among many other things, complex systems engineering, infrastructure design and future programming languages, cybersecurity assurance, and system verification.

AI cannot perform these tasks reliably, nor will it be able to in the foreseeable future. Human input remains essential, but pessimistic misinformation risks keeping tens of thousands of talented students away from important and meaningful careers in this vital field.

What AI can and cannot do

AI excels at making predictions. Generative AI improves on this by adding a user-friendly presentation layer to internet content: it rewrites, summarizes, and formats information to resemble human work.

However, current AI doesn’t really “think.” Instead, it relies on logical shortcuts, known as heuristics , that sacrifice accuracy in favor of speed. This means that, despite speaking like a person, it can’t reason, feel, care, or desire anything. It doesn’t work the same way the human mind does.

Not long ago, it seemed that ” prompt engineering ” (instructions, questions, or text) would replace IT. However, today, there are virtually no job openings for prompt engineers , while companies like LinkedIn report that the responsibilities of IT professionals have expanded.

The limitations of AI

What AI offers are more powerful tools for IT professionals to do their work. This means they can now take concepts further, from ideation to market implementation, while requiring fewer support roles and more technical leadership.

However, there are many areas where specialized human input remains essential, whether for reasons of trust, oversight, or the need for human creativity. Examples abound, but 10 areas stand out particularly:

1. Adapting a hedge fund algorithm to new economic conditions. This requires algorithmic design and deep market knowledge, not just reams of code.
2. Diagnose intermittent service outages in cloud providers like Google or Microsoft. AI can solve small-scale problems, but it can’t contextualize the resolution of large-scale, high-risk issues.
3. Rewriting code for quantum computers . AI can’t do this without extensive examples of successful implementations (which don’t currently exist).
4. Designing and securing a new cloud operating system . This involves high-level system architecture and rigorous testing that AI can’t perform.
5. Creating energy-efficient AI systems . AI can’t spontaneously invent lower-power GPU code or reinvent its own architecture.
6. Creating secure, cybercriminal-proof real-time control software for nuclear power plants . This requires combining knowledge of embedded systems with code translation and system design.
7. Verifying that a surgical robot’s software works under unpredictable conditions . Safety-critical validation is beyond the current scope of AI.
8. Designing systems to authenticate email sources and ensure integrity . This is a cryptographic and multidisciplinary challenge.
9. Audit and improve AI-based cancer prediction tools . This requires human oversight and ongoing system validation.
10. Creating the next generation of safe and controllable AI . The evolution toward safer AI cannot be the work of AI itself; it is the responsibility of humans.

Why IT is still indispensable

One thing is certain: AI will reshape the way engineering and computer science are done. But what we’re seeing is a shift in working methods, not a total destruction of the field.

Whenever we face a completely new problem or complexity, AI alone is not sufficient for one simple reason: it relies entirely on past data. Therefore, maintaining AI, creating new platforms, and developing fields like trustworthy AI and AI governance all require computing.

The only scenario in which we could do without computing would be if we reached a point where we no longer expected new languages, systems, tools, or future challenges. This is highly unlikely.

Some argue that AI could eventually perform all of these tasks. It’s not impossible, but even if AI were to become that advanced, almost all professions would be at the same risk. One of the few exceptions would be those who build, control, and develop AI.

There is a historical precedent: during the Industrial Revolution, factory workers were displaced at a ratio of 50 to 1 as a result of rapid advances in machinery and technology. In that case, the workforce grew with the new economy, but most of the new workers were those who could operate or repair machines, develop new machines, or design new factories and processes around machinery.

During this period of great change, technical skills were the most in-demand, not the least. Today, a parallel situation exists: technical knowledge, especially in computer science, is more valuable than ever.

Let’s not confuse the new generations with the opposite message.

Author Bio: Ikhlaq Sidhu is Dean of IE School of Science and Technology at IE University