An MIT Student Awed Top Economists With His AI Study—Then It All Fell Apart

https://www.msn.com/en-us/money/careersandeducation/an-mit-student-awed-top-economists-with-his-ai-study-then-it-all-fell-apart/ar-AA1QV7Rk

The preprint alone has been cited nearly 80 times. The student claims the company really existed but that the data was hard to access properly. Others believe he invented the whole study out of whole cloth.

 

The conclusion is — the paper cannot be trusted.

Indeed concerns were raised. A “computer scientist with experience in materials science” questioned how the data had been obtained, and how a lab — unknown in the public materials-science community — could produce such gains.

So as a result, MIT requested that the preprint be withdrawn from arXiv and that the submission to the leading journal The Quarterly Journal of Economics be withdrawn.

In short: as of now, the study is effectively retracted / disavowed by MIT and should not be relied upon for serious academic or policy arguments.

MIT hasn’t disclosed the full investigative details, there remains some uncertainty: we don’t know exactly what was fabricated, or who — if anyone beyond the author — was aware or complicit.

https://economics.mit.edu/news/assuring-accurate-research-record?utm_source=chatgpt.com