The Massachusetts Institute of Technology (MIT) has publicly disavowed a high-profile research paper that claimed AI boosts scientific output but lowers researcher satisfaction.
The study, “Artificial Intelligence, Scientific Discovery, and Product Innovation,” was authored by Aidan Toner-Rodgers.
He was a former doctoral student in MIT’s economics program. And his research gained national attention last year for its claims about AI’s role in innovation.
Now, MIT states that due to concerns about the “integrity” of the research, the paper should be removed from public discourse.
Details of the Study
The paper claimed that when an unnamed materials science lab used an AI tool, the lab produced more research discoveries and patent filings.
However, researchers reported lower job satisfaction.
The findings suggested that while AI increases productivity, it may harm workplace morale.
From Praise to Doubt
Two leading MIT economists initially supported the paper. Daron Acemoglu, a recent Nobel laureate, and David Autor, a respected labor economist, spoke highly of the research last year.
In an interview with The Wall Street Journal, Autor said he was “floored” by its insights.
That support has now changed.
In MIT’s latest statement, both economists expressed doubt. They wrote they have “no confidence in the provenance, reliability, or validity of the data and in the veracity of the research.”
The Investigation
This change in opinion followed a report from a computer scientist with expertise in materials science.
The person raised concerns about the paper’s data earlier this year. Those concerns were brought to MIT, which initiated an internal review.
Due to student privacy laws, MIT did not disclose the review’s findings. However, it confirmed the author is “no longer at MIT.”
In addition, the university requested the paper’s removal from The Quarterly Journal of Economics.MIT also asked for its withdrawal from the public preprint server arXiv.
However, according to arXiv policy, only authors can withdraw their papers. MIT noted that the author has not submitted a removal request to date.
