A scandal is shaking the academic community, with Nicholas Polson, a professor of econometrics and statistics at the University of Chicago Booth School of Business, at its center. During 2026, Polson was listed as an author or co-author on no fewer than 258 papers uploaded to the Social Science Research Network (SSRN), with artificial intelligence used in their preparation.
This means that, over a period of less than eight months, Polson put his name on an average of more than one paper per day. In August alone, 104 papers were uploaded, including six on a single day. These were not merely short notes or outlines for future research. Many ran to 30, 50, or even 80 pages.
It was not only the number of papers that attracted attention, but also the variety of subjects they covered. Alongside statistics, econometrics, financial markets, and machine learning, Polson’s name appeared on papers dealing with genetics, chess, psychology, and human relationships. 
Producing so many papers in such a short period would not have been possible without substantial assistance from artificial intelligence or collaborators. This raised an obvious question: how much of the work was written by the researchers, and how much by artificial intelligence?
Who Actually Wrote the Papers?
Polson did not write all 258 papers on his own. Vadim Sokolov, a professor of systems engineering at George Mason University, was listed as a co-author on more than 100 of them, while other names appeared on some of the papers as well.
That, however, does not make the case any less serious. Polson put his name on every paper on the list and thereby accepted responsibility for its contents. The same applies to the other named co-authors.
The problem is that it remains unclear who did what. There is no public information about which sections Polson wrote, how much his collaborators contributed, who checked the claims and calculations, or the extent to which artificial intelligence was involved in writing the papers.
Sokolov responded only after the case had attracted the attention of the academic community. He rejected the suggestion that the authors had simply entered a single prompt into an AI system and uploaded the resulting text without reading it. According to Sokolov, many of the papers were based on ideas, notes, mathematical work, computer code, lectures, and unfinished manuscripts that the authors had been developing for years. At the same time, he confirmed that AI had played a much greater role in the preparation of some of the newer papers and acknowledged that its use had not always been disclosed.
What Was Actually Published?
The papers in question were posted on SSRN, a platform where researchers share manuscripts before they undergo peer review and are accepted by academic journals. There is also no evidence that all 258 manuscripts were submitted to journals. Sokolov said that one of the papers had undergone several rounds of peer review before being published in CHANCE, a magazine published by the American Statistical Association.
Posting manuscripts that have not been peer-reviewed can nevertheless have consequences. They appear in search engines and academic databases, may be cited by other researchers, and can be incorporated into datasets used to train future artificial intelligence systems. The signature of a respected professor can, in itself, lend credibility to a paper.
From Statistics to Theories of Human Connection
Polson is not an obscure researcher seeking attention through unusual claims. He is an established statistician who has published notable work in Bayesian statistics, financial econometrics, and machine learning.
That is precisely why the case attracted so much attention. Andrew Gelman, a professor of statistics and political science at Columbia University, said that he knew Polson as a serious scholar but was taken aback by both the number and content of the new papers.
Gelman focused in particular on an 80-page manuscript titled Theories of Human Connection, which brings together economics, psychology, biology, game theory, philosophy, and spiritual traditions. Its authors attempted to combine these fields into a single model, presented in part through what they called a “master equation” of human relationships.
Gelman identified questionable assertions, inconsistencies in the data, and claims that were not supported by adequate scientific evidence. One concerned the purported ability to predict the breakdown of a relationship with an accuracy of more than 90 percent.
Sokolov later acknowledged that the claim required further qualification and said it would be corrected.
Papers Disappear Overnight
At the height of the controversy, the papers began disappearing from SSRN. Their number initially fell from 259 to 99, then to 44, then to just one. Eventually, the entire profile was removed.
Sokolov said that the papers had not been withdrawn by the authors but removed by the platform itself. According to him, SSRN told them only that it could not disclose the specific reasons for the removal.
Why the profile was taken down remains unknown. SSRN has not said whether the decision was prompted by the unusually large number of papers, the way artificial intelligence was used in their preparation, or some other reason. The University of Chicago has not yet commented publicly on Polson’s case.
Photo> ChatGPT
