Jukka Suomela ยท June 2, 2026
On May 28, 2026, we organized a discussion event with our faculty members, with the focus on these questions:
These notes summarize some of the key themes of the discussions. Many thanks to all participants!
Entry-level employees are expected to manage armies of AI assistants. They need skills and competences in management, organization, delegation, exception handling, accountability, and quality control.
We are living in a new, increasingly complex world, with new demands, and a new kind of mental load. Long-term use of AI tools can have a serious impact on our well-being and cognitive skills.
Our students need to understand how learning works, what the use of AI tools can do to their brains, and how all the new tools can also help us with learning if used in the right way. Our students need a systematic engineering attitude to learning new things. Technology evolves rapidly, and there will be a need to learn quickly and pivot quickly.
Our students still need to learn how to work with human beings. They need to have opportunities for having conflicts and disagreements in a civilized way, and they need to also learn to fail.
There seems to be broad consensus that experts benefit most from AI. It is even more important than ever before that our graduates are top talents. Our students need to know how to become an expert.
Strong oral and written presentation skills are essential. Our students need to be able to read long texts. They need to be able to think, explain their thinking clearly, and have patience to take time with solving problems. They need to be able to identify, judge, and produce sound reasoning.
Projects are getting bigger and more complicated. It is essential to be able to organize documentation, do strategic planning, and manage and design complex systems; handover of complex systems will need special attention.
Our students will need broad knowledge of all areas of CS, and beyond. Companies need “human Swiss Army knives.” One-person teams can nowadays create and deploy large systems. The same person has to understand computing clusters, operating systems, databases, compilers, algorithms, machine learning, computer networks, user interfaces, testing, software engineering, deployment, operations, information security, customers, business, legal issues, ethics, and a lot more. Students need to learn how to create their own business, and how to make it AI-proof.
The volume of AI-generated output far exceeds the bandwidth of human verification; we need rigor and formal methods to make extensive use of AI effective even in critical systems. The key skill is writing specifications. Foundational competences include mathematical proofs, formalization of mathematics, formal verification, and proof assistants. Our students need to learn about fail-safes, access control, safety, and security. They need to understand what production quality really means in different contexts.
We already know a lot about AI, but there is a lot to be discovered by future researchers. It is our task to educate the next generation of CS researchers who will make the future breakthroughs related to AI.
The working methods of exact sciences and bottom-up reasoning from first principles alone are unlikely to get us there. Our students will also need the working methods of empirical sciences. They need tools for understanding extremely complex systems and emergent behavior in such systems. They need to understand how new scientific knowledge is created in the first place.
CS is suddenly at the core of big societal, political, environmental, ethical, legal, and financial issues. Our students need to understand complex global issues that go far beyond our traditional comfort zone. It is critical that they also understand their own professional and ethical role and responsibility.
It is important to understand the difference between delegating up (using AI to solve something that is beyond your own capabilities) vs. delegating down (using AI to automate mundane work that you could do by yourself); these are distinct skills and need different approaches.
Our graduates need to be able to identify the right problems to tackle, find the right solution strategies, and be aware of tools and their limitations and costs. They need to understand the legal framework that governs AI usage.
How does AI impact what we expect from our incoming students in the future? Do we need to adjust our admission criteria? Conversely, what does our degree certificate certify? How does a university graduate fundamentally differ from someone who has self-studied with the help of AI?
We need to focus also on how to teach, not just what to teach. To keep the discussion focused, we need to separate teaching from assessment.