Last year, the Massachusetts Institute of Technology (MIT) was enthusiastically promoting a PhD student’s research on how artificial intelligence is transforming the workforce—work that had “floored” seasoned professors in the field. But now, the university is distancing itself from the study and requesting that it no longer be publicly available.
On Friday, MIT announced it had conducted a review of the research following concerns raised about its validity. The outcome: the paper should be “withdrawn from public discourse.” Titled Artificial Intelligence, Scientific Discovery, and Product Innovation, the study attracted widespread attention for its claim that scientists using AI tools were significantly more productive than those without them. However, it also noted a downside—those same scientists reported lower satisfaction with their work.
The study was initially hailed as groundbreaking. Nobel Prize-winning economist and MIT professor Daron Acemoglu called it “fantastic.” But not everyone was convinced. According to the Wall Street Journal, a computer scientist with expertise in materials science questioned the AI tool used in the research—specifically how it functioned and whether it truly drove the claimed surge in innovation.
Those concerns were brought to other MIT faculty members, prompting a formal review. The university ultimately concluded it “has no confidence in the provenance, reliability or validity of the data and has no confidence in the veracity of the research contained in the paper.” MIT did not detail what specific issues were uncovered, citing student privacy laws and internal policies.
The author of the study is no longer affiliated with MIT, and the university has requested that the preprint be removed from arXiv. It has also withdrawn the paper from submission to the Quarterly Journal of Economics, where it had been under review for possible publication.
David Autor, an MIT economist who previously praised the study, told the Wall Street Journal, “More than just embarrassing, it’s heartbreaking.” The reversal represents a significant setback for research exploring AI’s role in scientific innovation. What once appeared to signal a wave of AI-driven breakthroughs is now under serious doubt—raising broader questions about how AI really impacts discovery, and how much we can trust early findings in this fast-moving field.
