MIT has called for the withdrawal of a widely discussed research paper on artificial intelligence, citing serious concerns about the paper’s data integrity and reliability. The study, titled “Artificial Intelligence, Scientific Discovery, and Product Innovation,” was authored by a doctoral student in MIT’s economics program. It claimed that introducing an AI tool into a large materials science lab led to increased material discoveries and patent filings, but also decreased researchers’ job satisfaction.
Renowned MIT economists Daron Acemoglu and David Autor had initially praised the study, with Autor describing it as “flooring.” However, they now say they have “no confidence” in the accuracy of the data or the validity of the findings. The shift came after a computer scientist raised concerns in January, prompting MIT to launch an internal investigation.
While MIT hasn’t revealed the outcome of that review due to student privacy rules, it confirmed that the student is no longer affiliated with the university. The paper’s author has been identified in earlier coverage as Aidan Toner-Rodgers.
MIT has also requested the paper be withdrawn from both The Quarterly Journal of Economics and the preprint server arXiv. So far, the author has not submitted the withdrawal to arXiv, despite the university’s request.
