Quad and Cloister
Live
Institutions

Dartmouth Provost Resigns Over Undisclosed AI Writing

Dartmouth University Provost Santiago Schnell resigned after a faculty committee found he used AI to substantially edit academic and public writing without disclosure, violating integrity standards ahead of a no-confidence vote.

Dartmouth University Provost Santiago Schnell resigned after a faculty committee found he used AI to substantially edit...

Santiago Schnell resigned as Dartmouth's provost and professor of math and biology on Wednesday. His departure came less than two weeks before faculty leaders planned a no-confidence vote, following a one-month independent review into his undisclosed use of artificial intelligence in his writing.

President Sian Leah Beilock announced the resignation, which followed investigations by a news outlet and the student newspaper. The faculty-led committee, formed on September 25, submitted its report on October 27. The findings exposed a stark tension at a university publicly championing AI adoption while grappling with its ethical application.

AI use and disclosure failures

The committee documented extensive AI use in Schnell's academic and public writing without the required disclosures. In one journal paper published in September in Chemistry Europe, Schnell used Claude to substantially edit his research manuscript, troubleshoot computer code, and generate figures. He provided no disclosure.

When the journal's editor later inquired, Schnell said he did not understand that this type of AI-assisted editing required disclosure. The committee found his description understated his actual AI use. In a forthcoming article, he uploaded his manuscript to Claude, asked if it contained AI writing, and then submitted a version with Claude's suggested changes.

The review found the majority of Schnell’s recent written work, including academic articles and newspaper op-eds, was generated by AI and published without disclosure. He used AI tools for multiple tasks across various publications.

Schnell did not fully cooperate with the investigation, failing to provide logs from his personal AI accounts upon request. He told Beilock in a September email that he relied on older author instructions but later acknowledged a disclosure would have been the best approach. The committee concluded his actions violated the policies of publishers and a national newspaper, falling short of Dartmouth's standards for ethics and professionalism. A separate analysis by Pangram determined one of his Washington Post editorials was at least 80 percent AI-generated.

Committee findings and institutional response

The committee, led by former provost and computer science professor David Kotz, concluded Schnell violated Dartmouth's academic integrity and honesty standards. Members included Hany Farid, Peter Orner, Lindsay Whaley, and Odette Harris. They did not use AI detection tools like Pangram, instead reviewing Schnell's logs from Dartmouth's enterprise AI tool Claude and his emails.

Despite finding violations of the university's Code of Ethical Conduct and Academic Honor Principle, the committee did not recommend any disciplinary action. Schnell has published at least 30 pieces since joining Dartmouth in July 2025. The committee found no violation in a fourth paper he co-authored.

President Sian Leah Beilock stated leaders must model transparency and human judgment. "I expect our senior leaders to hold themselves to the highest standards," she said. She declined to say whether Schnell’s use of AI was right or wrong. Faculty and students noted the contradiction between Schnell’s actions and Dartmouth’s public embrace of AI, which includes a contract to give all students access to Claude.

Creative writing professor Jeff Sharlet said Schnell’s resignation was a good day for teachers and students who value clear academic standards. Senior Ben Stevenson said the situation highlighted a clear question of academic integrity and that faculty and student pushback forced the provost’s hand.

The university is now preparing to host a three-day conference titled 'AI at 70: What Must Remain Human?' starting Thursday, celebrating the 70th anniversary of the 1956 conference where Dartmouth claims to have coined the term 'artificial intelligence.' The committee's report, dated October 27, lands as the institution publicly examines the very technology that precipitated the crisis.

Related coverage

More from Institutions