GoKawiilThe mathematics community is right to call for transparency, integrity and fairness to be protected when AI tools are used. Researchers in other disciplines could learn from this approach.
The city of Leiden in the Netherlands is gaining a reputation for hosting meetings on integrity in science.Credit: Getty
A little more than a decade ago, after a 2014 conference at Leiden University in the Netherlands, researchers published the Leiden Manifesto in Nature. It called for the responsible use of metrics in research and included a set of ten principles to help to ensure rigour and fairness in the evaluation of research1. Along with the San Francisco Declaration on Research Assessment (DORA), the Leiden principles have been adopted around the world.
How AI is reshaping discovery in maths and physics
The principles came about in response to a suite of metrics data and innovative computational tools in universities and science that required guardrails for their use. Last September, the city of Leiden and its university reprised their part as a meeting point for scholars concerned about new technologies and the integrity of science — this time, about the role of artificial intelligence in mathematics.
The resulting Leiden Declaration on Artificial Intelligence and Mathematics, published earlier this month2, shares some of the concerns of its namesake. It recognizes the power and potential of a transformative technology while urging researchers and institutions to ensure that human judgement, transparency and fairness are protected — principles that are foundational to science and must remain so. The declaration has been gaining endorsements from researchers across the discipline, which includes those who are deeply sceptical of AI and those who are much more optimistic. Nature wholeheartedly endorses both the declaration process and its conclusions.
The Leiden declaration and the preceding workshop brought together researchers from maths, computer science, philosophy and history. AI is transforming learning and research in the field, as a group of London-based mathematicians wrote in a Nature Comment last week3. Applications range from automating the checking and verification of mathematical proofs to helping to solve — or even autonomously solving — open problems in particular areas of maths. Only last month, an 80-year-old challenge in geometry, known as the unit-distance problem, was solved by mathematicians at the US technology firm OpenAI using only a single prompt to a chatbot.
Why we have nothing to fear from the decolonization of mathematics
In the words of Fields medallist Terence Tao at the University of California, Los Angeles, AI is rapidly changing mathematicians’ job descriptions. But it is arguably doing more than that. As AI and commercial AI software become integrated into maths research, they “will change the kinds of problems that are pursued and the forms of proof that are valued”, according to a report from the workshop4. The Leiden declaration warns of the risks this poses to the autonomy of the field. Indeed, evidence is starting to emerge that, across science, the use of AI correlates with a narrower breadth of research topics5. As the declaration states, this will disadvantage both researchers who lack access to the technology and those who do not wish to use proprietary AI tools.
The Leiden declaration holds that mathematical results should continue to be published in peer-reviewed venues subject to the principles of open science, and that “no proprietary knowledge or equipment should be required to understand them”. Moreover, any material used as training data should have the required attributions and should not be used without consent.
... continue reading