A group of 16 mathematicians, working with colleagues and mathematical organizations worldwide, last week published the Leiden Declaration on Artificial Intelligence and Mathematics, a set of guidelines aimed at protecting the discipline's core values.

The declaration urges governments and funding bodies not to "believe the hype" surrounding AI systems' mathematical capabilities, according to science news site Phys.org.

The move comes after AI models began making headlines for tackling research-level mathematics. In late May, OpenAI, the maker of ChatGPT, announced that one of its models had disproved a notable 80-year-old conjecture in combinatorial geometry, according to The New York Times. The conjecture was one of about 1,200 open questions proposed by Hungarian mathematician Paul Erdos, widely regarded as one of the most prolific mathematicians in history.

Mathematics problems on a blackboard. Photo from Pexels

Alongside its research paper, OpenAI released a companion assessment by several independent mathematicians. "No previous AI-generated proof has come close" to meeting that standard, British mathematician Timothy Gowers, winner of the 1998 Fields Medal, an award often described as the mathematics equivalent of the Nobel Prize, wrote in commentary shared by the company.

The declaration's authors said they are not opposed to AI itself but are concerned about how commercial developers present and use the technology.

According to the statement, companies have strong financial incentives to promote AI breakthroughs while revealing little about their methods, training data or unsuccessful attempts. The authors argued that mathematical research should remain open, transparent and independently verifiable.

They also warned that AI-generated proofs can appear convincing while containing subtle errors that may be difficult for journal editors and reviewers to detect.

The declaration calls on researchers and AI developers to disclose the use of AI in mathematical work, properly credit earlier research, preserve rigorous peer review and ensure that results can be independently checked. It also advocates stronger public support for academic mathematics to prevent the field from becoming overly dependent on large technology companies.

Rodrigo Ochigame, one of the declaration's authors, said mathematicians who "never intended to contribute to AI development" are increasingly seeing their work used to train commercial models without their consent, Scientific American magazine reported. "I think that’s a deeply concerning situation," he added.

The declaration has received the backing of the International Mathematical Union.

In an endorsement, IMU Vice President Ulrike Tillmann said AI "opens new and exciting opportunities, but it also raises questions that cannot be left unexamined."

"The future of mathematical research must be guided by human judgment, fair and transparent practices, and the shared values of the global mathematical community."