Researchers Unite to Promote Responsible Use of AI: Scientific Autonomy at Stake

In June 2026, approximately 60 experts from ten countries—including research mathematicians, computer scientists, philosophers, and social scientists—issued a formal statement on artificial intelligence and mathematics following eight months of consultation. This community initiative, officially known as the Leiden Declaration on AI and Mathematics, has already been endorsed by some of the world’s most prestigious professional organizations, including the International Mathematical Union (IMU), as well as prominent scientists such as Fields Medalist Terence Tao. Several senior researchers at the HUN-REN Alfréd Rényi Institute of Mathematics have also signed the Declaration.

“What emerged was not a compromise, but a shared position based on common ground. It is neither a belated reaction nor a cry for help; rather, we intend it as a wake-up call,” emphasized Bryna Kra, Professor at Northwestern University, former President of the American Mathematical Society, and a member of the preparatory working group, in an interview published on the HUN-REN Rényi Institute’s website. (The full text of the interview is available here.) “The time has come for mathematicians to engage with these issues because our field is changing rapidly, and if we want a say in what it becomes, we can no longer remain silent. (...) Many of the loudest voices expressing opinions are not mathematicians at all, but predominantly representatives of companies that have investments and therefore commercial interests.”

Scientific Integrity Under Threat

The Declaration stresses that mathematics is not simply about the automated generation of theorems, but about human understanding and the transparency of proofs. However, the rapid development of artificial intelligence—from systems capable of solving International Mathematical Olympiad problems to models tackling challenges at the PhD level—directly threatens the core values of the discipline:

• Proof and Understanding: Mathematical proofs confer the highest degree of certainty and impart an understanding of why conclusions are true. Current artificial intelligence technologies generate arguments that appear plausible yet may remain unreliable or incorrect. These outputs are difficult to distinguish from correct mathematical proofs, placing the present system of review under increasing pressure and jeopardizing traditional standards of correctness.
• Authorship and Responsibility: Scientific results must remain attributable to specific authors who take credit for their discovery and assume responsibility for their correctness. Models trained on the published mathematical commons frequently synthesize knowledge without properly citing the human works they utilize, often systematically exploiting licenses or violating copyright protections.
• Transparency and Verifiability: Mathematical arguments must be transparent and subject to independent verification. In principle, understanding them should require no proprietary knowledge or equipment.
• Autonomy: A key source of strength for the discipline has long been the autonomous shaping of the direction of research and the methods used to pursue it. This autonomy is increasingly threatened by the growing involvement of technology companies in mathematical research. This trend raises the risk that research questions may come to be prioritized because of their amenability to automated mathematics, rather than expert judgment of their deeper scientific significance.

Recommendations and Ethical Responsibility

The Declaration does not call for the rejection of artificial intelligence; rather, it advocates its responsible deployment. Researchers are urged to transparently disclose the use of automated tools and computational resources in their publications, while the document clarifies that the responsibility for correctness and the accuracy of citations remains exclusively with the human authors.

The statement also calls for rigorous ethical reflection. Because mathematics leads to technologies used in warfare, mass surveillance, and the undermining of democracy, the authors advise researchers to evaluate the ethical consequences of their research and, where necessary, withdraw from harmful work.

A Model for Other Disciplines

The significance of the Leiden Declaration extends far beyond the boundaries of mathematics. Its message applies equally to education, law, and the wider scientific community, where artificial intelligence risks creating highly asymmetric terms between academia and technology giants. As Professor Bryna Kra pointed out, future generations will carry systems in their pockets capable of answering questions on their behalf; consequently, the role of explanation and human understanding must be redefined immediately.

As a continuation of the national dialogue, the Section of Mathematics of the Hungarian Academy of Sciences (MTA), the Rényi Institute, and the János Bolyai Mathematical Society are organising a joint event this autumn to discuss local guidelines inspired by the Declaration.