Dutch mathematical society takes action on AI
20 November 2026
The Koninklijk Wiskundig Genootschap (KWG), the Dutch Royal Mathematical Society, has adopted the Leiden Declaration on Artificial Intelligence and Mathematics. Its General Assembly also gave KWG a mandate to draft concrete recommendations in response to the groundbreaking effects of AI on mathematical research, education and careers.
These decisions were taken at the end of the one-day event “AI in Mathematics”, held at the Utrecht Science Park on 23 October. More than 150 mathematicians based in the Netherlands, working both in academia and in industry, attended the gathering, which included five talks and an extended panel discussion. The speakers voiced a variety of viewpoints, ranging from practical recommendations about effective AI use in research and in teaching, to ethical and cultural considerations, both within the mathematical community and within society as a whole.
On the one hand, the speakers demonstrated impressive applications of AI in a diverse range of mathematical disciplines, real-world applications, and teaching. Examples included learning guided by chatbots, AI-assisted formalization of mathematical results, and enabling the larger public to contribute to mathematical research via large collaborative projects whose correctness is guaranteed by formalization. Many shared the feeling that, in the hands of researchers with the necessary expertise, AI has enormous potential to accelerate progress.
At the same time, the discussions also revealed concerns about the scale and speed at which AI is changing the mathematical research landscape. Unreleased state-of-the-art AI models have recently solved high-profile mathematical questions that had been open for decades. This is very exciting, but has also clashed with many of the values that mathematicians hold dear. Indeed, for many of us, beauty is at the core of what we do. We seek for better, more elegant explanations of mathematical phenomena. This labor of synthesis and abstraction, building on the work of our predecessors, has allowed us to conceptualize increasingly complicated ideas and solve problems that were formerly out of reach. Ideas that were groundbreaking for XIX century mathematicians are now part of elegant, streamlined theories that can be taught to undergraduate students. When AI solves a major problem, it is standing on the shoulders of generations of mathematicians.
This realisation has deep repercussions, particularly when it comes to Mathematics education. This was one of the most discussed themes during the meeting, and possibly the one that needs to be most urgently addressed. Mathematics departments train students at the bachelor, master, and PhD levels, who then go on to contribute to industry, society, and academia. There is a very real risk that overreliance on AI may allow students to bypass the effort required to develop critical reasoning skills, leading to a loss of talent. Doctoral training raises further questions: How should we evaluate a candidate’s contributions? What are suitable criteria to assess the mathematical excellence of a PhD trajectory? How can we plan research projects three or four years into the future, given the current speed of progress? These are dilemmas that hiring committees will face very soon; the careers of young mathematicians depend on us being able to provide clear answers.
KWG’s mandate, approved at its General Assembly, is to respond to these concerns. It is strongly felt that the mathematical community needs to take action to ensure that technological progress supports mathematical understanding and the people who develop it. KWG, in close connection with Platform Wiskunde Nederland, NWO, and various other stakeholders, will now work towards concrete recommendations within the context of the Dutch mathematical and scientific landscape.