MIT has disavowed a PhD student's research on the impact of AI on the workforce and is taking steps to have it withdrawn from public discourse. The university formally requested that the preprint, which had garnered attention from field experts and economists, be retracted from arXiv and The Quarterly Journal of Economics.

The paper, titled "Artificial Intelligence, Scientific Discovery, and Product Innovation," was published on arXiv in November 2024. It claimed that the introduction of AI in a materials science lab led to increased material discovery and patent filings but reduced researcher satisfaction. Despite being a preprint that had not undergone peer review, the study was acknowledged by MIT economists Daron Acemoglu and David Autor, who initially expressed enthusiasm for its findings.

However, after the paper's release, Acemoglu and Autor found inconsistencies in the study and brought their concerns to MIT. Following an internal review, MIT decided the paper should be withdrawn and requested the author to submit a formal withdrawal request. When the author did not comply, MIT directly requested arXiv to mark the paper as withdrawn.

Acemoglu and Autor released a joint statement alongside MIT's announcement, explaining that they had concerns about the validity of the research and brought it to the attention of MIT. The researcher responsible for the study is no longer affiliated with the university.