Dartmouth provost resigns after review finds undisclosed use of AI

Santiago Schnell, dean of the College of Science at the University of Notre Dame in Indiana, has been named Dartmouth’s next dean, the college announced Tuesday. Photo credit: University of Notre Dame – Barbara Johnston

This Molly Hughes story was first published in the Valley News on October 28, 2026.

HANOVER — Dartmouth Provost Santiago Schnell has resigned and is leaving the university after a faculty-led committee found he used artificial intelligence in his writing without properly disclosing it.

President Sian Leah Beilock announced her resignation Wednesday afternoon in an email to the campus community.

“Out of respect for his colleagues and the institution, Santiago has resigned as dean and is leaving the university,” Beilock wrote. “It was his decision.”

Beilock formed the committee on September 25, after news reports questioned Schnell’s use of AI, including an August Semafor article on AI-written opinion pieces that named his Washington Post column.

In its report, dated October 27, the committee found that Schnell “used AI to significantly edit his writings in violation of at least one AI use and/or disclosure requirement from at least one publication, failed to disclose his use of AI in clear violation of the stated policy of a national newspaper, was not open about his use of AI” and failed to meet the standards of ethics and professionalism of Dartmouth.

The committee said it does not use AI detection tools. Instead, he reviewed Schnell’s logs from Dartmouth’s enterprise AI tool, Claude, as well as his emails. He also said he was not recommending any disciplinary action.

Schnell has published at least 30 articles since joining Dartmouth in July 2025, according to the report. Given the short time frame, the committee focused on a handful of them.

In a newspaper article, the committee found that Schnell used Claude to “extensively edit his research manuscript, resolve errors in the computer code, and generate figures.” The newspaper required that this type of use be disclosed, and Schnell did not disclose it. He told Beilock in a September 22 email that he had relied on instructions from an older author.

“This was not the case, but disclosure would have been the best approach,” Schnell wrote, according to the report.

The committee said the publisher has required AI disclosure since at least 2023.

Schnell told Semafor that he used ChatGPT “to refine some arguments, check for grammatical errors and submit them to the Washington Post.” The committee said it was a “blatant violation” of Post policy.

According to the committee, in an upcoming article, Schnell uploaded his manuscript to Claude and asked, “Are there many AI signatures in this article?” He then submitted a version with the types of changes Claude suggested. When the journal’s editor-in-chief asked about the use of AI, Schnell said he “did not understand that this type of AI-assisted editing required disclosure.” The committee said its description “understates the nature of the Provost’s actual use of AI.”

The committee found no violations in a fourth article, co-authored by Schnell.

The report also states that Schnell did not fully cooperate.

“Although he provided his Dartmouth-issued laptop for inspection, the Provost failed, upon request, to provide logs of his use of the personal LLM accounts that he used for his academic and public writings,” the committee wrote.

Former dean and computer science professor David Kotz led the committee. Its other members were computer science professor Hany Farid, English professor Peter Orner, linguistics professor Lindsay Whaley, and administrator Odette Harris.

Schnell did not respond to an email seeking comment Wednesday, and Jana Barnello, Dartmouth’s director of media relations and communications strategy, did not immediately respond to a request for comment.

History professor Matthew Delmont will serve as interim dean until June 2028.

In his letter, Beilock said leaders must “set an example of the transparency and humane judgment we expect from our students and scholars.” Dartmouth is organizing its “AI at 70: What should remain human?” » conference later this week.