Mathematicians warn of AI threats to profession as industry encroaches (arstechnica.com)

Mathematicians have raised alarms about the influence of artificial intelligence on their field, culminating in the Leiden Declaration on Artificial Intelligence and Mathematics. This declaration, developed by a group of 16 researchers over eight months and endorsed by the International Mathematical Union, highlights the potential threats posed by AI technologies, especially following recent claims from OpenAI about an AI model disproving an 80-year-old geometric conjecture. The timing underscores growing concerns regarding how these advancements could undermine the integrity and values still held by the mathematics community. The declaration warns that AI-generated drafts can present plausible but incorrect arguments, complicating the traditional standards of correctness and transparency in mathematical proofs. Mathematicians face increased pressure to review and validate these AI-generated outputs, which could lead to a proliferation of erroneous literature and flawed results. Notably, the document emphasizes that this trend disproportionately affects students and early-career mathematicians, ultimately endangering the discipline's future. As Leslie Ann Goldberg of the University of Oxford noted, erroneous AI contributions may clutter academic literature, undermining the foundation on which new research builds and potentially eroding confidence in mathematical rigor.