A group of twenty-five distinguished mathematicians has released an open letter expressing concerns that artificial intelligence laboratories are undermining their intellectual contributions as they compete to solve renowned mathematical problems. Each signatory is a recipient of the Fields Medal, widely regarded as the highest honor in mathematics.
Recently, Tristan Buckmaster, a professor at NYU, accused OpenAI of attempting to influence him against acknowledging a collaborator associated with Anthropic regarding a significant mathematical discovery. He questioned whether the firm used insights from their Codex tool to create an innovative proof over an intensive weekend of analysis.
On Thursday, OpenAI retracted its sponsorship of an upcoming mathematics event at CalTech following backlash from university researchers.
The mathematicians who signed the letter assert that while AI’s capability to tackle fundamental mathematics could greatly benefit society, this potential hinges on whether these solutions can be effectively understood and shared within the mathematical community and beyond.
They noted, “These solutions are often presented hastily, lacking the necessary time for comprehensive documentation, isolation of new techniques, and acknowledgment of prior relevant work.” They also pointed out that OpenAI’s proof is still unverified. “This situation raises significant concerns regarding attribution and plagiarism. Furthermore, without the active involvement of mathematicians to develop and integrate these AI-generated ideas into the mathematical framework, the essential communication chain among mathematicians would be jeopardized.”
Amid fears of their work feeding into OpenAI’s models without proper credit, many mathematicians are increasingly uneasy. Currently, if leading labs identify a promising avenue for discovery, they can invest substantial resources into using large language models to outpace the original researchers in proving concepts—promoting a culture of secrecy.
This open letter echoes sentiments expressed in the Leiden Declaration, issued by a group of mathematicians in June, which also addresses the implications of AI-generated proofs on their work and provides recommendations for mathematicians, academic institutions, and policymakers.
Similar to other fields undergoing transformations due to AI, the mathematical community argues that the essence of mathematics extends beyond the proofs and attribution of credit. It encompasses the intellectual groundwork that supports students, generates new inquiries and ideas, and integrates them into the wider tapestry of human knowledge.
If you think the issues facing the high-stakes world of mathematics are isolated, bear in mind that your area of interest may soon encounter analogous challenges.
The mathematicians stated, “The dilemmas currently confronting the mathematical field mirror those affecting many scientific and creative disciplines, underscoring a collective challenge humanity must address: ensuring that as AI reshapes our work processes, we do not overlook the fundamental goals these efforts were originally intended to achieve.”
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.
