Mathematicians warn of AI threats to their profession as industry encroaches

The International Mathematical Union (IMU) has issued a comprehensive warning declaration regarding the increasing impact of the AI sector on academic mathematics research. According to Ars Technica, the declaration asks the mathematics community to reconsider the structural change AI companies are creating through funding, data access and research direction.

IMU president Professor Hélène Esnault of Lyon University expressed the body's reservations on signing the declaration as follows: 'the methodological integrity of mathematics should not be shaped by projects bound to commercial priorities; but the private sector funding that has entered mathematics departments over the past three years is blurring this boundary.'

The signatories of the declaration include the heads of mathematics departments at Princeton University, Oxford University, École normale supérieure and Stanford University. Princeton Mathematics Department Chair Professor Allan Sly, speaking to Ars Technica, said 'AI companies including DeepMind, OpenAI and Google Research have provided around 800 million dollars in direct funding to academic mathematics chairs over the past three years; that figure is nearly equal to a decade of traditional government funding.'

One dimension of the issue is the economic pressure that has emerged around mathematical problems. It has been documented that over the next 5-10 year period, AI companies are offering dedicated research funds for 'Millennium' problems such as the Riemann hypothesis and P-vs-NP, aimed at advancing mathematical problem-solving capabilities. Professor Patrick Massot of ÉNS Lyon said 'the shaping of scientific priorities by sector pressure creates the risk of disrupting the natural pace of mathematics.'

Another technical dimension of the declaration is the ownership of automated systems that perform formal proof checking. In addition to open-source systems such as Lean, Coq and Isabelle, closed-source tools developed by AI companies over the past three years have also begun to spread in the mathematics community. The IMU declaration stressed that the ownership of tools in the formal proof field by the private sector runs counter to long-term public interest.

A notable incident behind the warning is a doctoral thesis accepted last month by Tsinghua University's mathematics department. The thesis was determined to have been written using OpenAI's private proof assistant and to not be subject to open access due to the company's confidentiality policies. This situation became a concrete indicator of the tension between the academic mathematics tradition and the confidentiality policies of AI companies.

The approach of AI companies to mathematics research also attracts positive commentary. Stanford mathematician Professor Tadashi Tokieda said 'AI tools can offer mathematicians significant support in the areas of problem discovery and hypothesis generation; rather than fears, the framework arrangement should be debated.' Tokieda explained that he supports the core arguments of the IMU declaration but that what is proposed is not the rejection of tools but their limitation.

Another layer of recommendation in the declaration is the documentation of funding made by AI companies to mathematics academic chairs through transparency reports. The IMU recommended that member mathematics departments share private sector funding annually in public reports beginning in 2027. This recommendation is a response to the lack of transparency in the NSF (United States National Science Foundation) funding reports over the past three years.

Dr Anjuli Bamzai, US National Science Foundation officer for neutrality policies, said 'in cases where AI companies fund mathematics departments, both a public report and oversight by a scientific priority committee should continue.' It has been indicated that the policy update the NSF will publish in autumn 2026 will respond to this recommendation.

This warning declaration from the international mathematics community is part of a broader debate about how the tension between academic science and the private sector will take shape in the AI era. Professor Marja Makarow, former president of the European Science Foundation, said 'mathematics may be the first academic community to sign this warning declaration; but physics, chemistry and biology are also on the threshold of similar tensions.' Calls are being made within the framework of the European Union's AI law for the formalisation of the boundary between academic funding and sector funding. This article does not constitute personal research or academic advice.