Researchers Confide in AI, But Guard Their Data
A mathematician suspects an AI model used his unpublished research, raising concerns about data security for researchers. India is developing new AI regulations.
Mathematician Tristan Buckmaster from New York University suspects that an OpenAI AI model may have used his unpublished work to solve the Navier-Stokes equations. Buckmaster claims the AI produced a potential proof for a mathematical problem he had spent years on, in a matter of days.
The incident has sparked wider concern among researchers, engineers, and academics who routinely use AI tools for unpublished papers, code, and ideas. The worry centers on how researchers can safely leverage AI without exposing work that is not yet public.
In India, institutions are moving to address this. IIIT Hyderabad and MAHE Bengaluru have formed committees to draft institutional guidelines for the responsible and ethical use of AI, including secure handling of unpublished research.
Indian legislation currently lacks comprehensive safeguards for researchers in the age of AI. While attempts have been made to create rules, such as on the intersection of generative AI and copyright, they are not yet law and do not fully address researchers' interests, like the potential loss of priority and credit.