Researchers Raise Concerns Over AI Overreach and Data Privacy
Concerns are growing among researchers and academic institutions regarding the ethics of AI model usage and ownership of research findings, following incidents like OpenAI's Codex potentially solving the Navier-Stokes equations.
Academic circles are experiencing growing unease regarding the ethics of AI model usage and the ownership of research findings. The ability of OpenAI's Codex model to rapidly solve complex mathematical problems has raised questions about how training data is collected and utilized.
The incident involves mathematician Tristan Buckmaster, whose unpublished research was allegedly used to train OpenAI's model. Although OpenAI denied using user inputs for that specific solution, the situation exposed potential risks of losing academic credit and priority when confidential information is fed into AI tools.
In India, academic institutions are beginning to react. Many researchers use AI solely for literature reviews and avoid inputting complete solutions or patentable methods. Institutions are also considering open-source models and restricting the handling of sensitive data due to lack of control.
India currently lacks specific AI legislation for the research sector. Existing proposals focus more on financial compensation, but experts point out that for researchers, the loss of credit and credibility is also a significant harm. Until the situation clarifies, researchers should adhere to strict protocols, such as filing patents before AI use and ensuring agreements that prevent data from being used for training purposes.