The Pulse
Dartmouth Provost Santiago Schnell Resigns After AI Review
Dartmouth Provost Santiago Schnell resigned on October 28, 2026, after a faculty-led committee found undisclosed AI use in his writing. The committee identified publication-policy violations in a Washington Post column and an academic paper
AI.info Team · Written with AI · fact-checked ·
Dartmouth Provost Santiago Schnell resigned on October 28, 2026, after a faculty-led committee found that he used artificial intelligence to materially alter his writing and failed to meet publication disclosure rules in at least two cases. College President Sian Leah Beilock announced that Schnell was leaving the university and said the resignation was his decision. Matthew Delmont, a history professor and associate dean, will serve as interim provost through June 2028.
“Out of respect for his colleagues and the institution, Santiago has resigned as provost and is leaving the university,” Beilock wrote in her October 28 message to the Dartmouth community. The committee’s findings came from a review of Schnell’s academic publications and public writing, including his published work since joining Dartmouth in July 2025.
Two publication policies, two disclosure failures
The committee found that Schnell used Dartmouth’s Claude enterprise tool to substantially edit a research manuscript, troubleshoot computer code and generate figures for a paper in the journal ChemSystemsChem. He did not disclose that use when submitting or publishing the paper, according to the committee report. The report says the journal’s policy required disclosure at the time and that historical versions of Wiley’s guidance showed such a requirement had been in place since at least 2023.
Schnell told Beilock in a September 22 email that he had relied on older author instructions, the committee reported. He later disclosed using ChatGPT for language refinement, code review and consistency checks, but did not mention his use of Claude in that email to the journal. The committee said its review did not support his claim that Wiley’s disclosure requirement began in 2025.
The committee also found a clear violation of the Washington Post’s rules in Schnell’s August 9 guest column about AI cheating. In a statement to Semafor, Schnell said he used ChatGPT “to refine a few arguments, check for grammatical errors” before submitting the column. The Post’s policy required writers to confirm that submitted work had not been created or manipulated with AI or editing software.
The committee separated code assistance from writing
A third published work received a different finding. Schnell co-authored a paper in the Bulletin of Mathematical Biology with two junior academics, one of them a graduate student; the report says he later told his co-authors he had run their code repository through Claude Max Code and ChatGPT Pro Code for optimization. The journal’s policy changed between submission and publication, and the committee said the available evidence did not establish a policy violation.
The report also says the committee found no evidence that Schnell used AI to edit or manipulate the text of that paper. Its findings therefore distinguish the use of AI to review or optimize code from the use of AI to revise writing, rather than treating every use of the tools as the same conduct.
In a separate case involving a forthcoming article, the committee found that Schnell was not forthcoming about his AI use. It said he uploaded a manuscript to Dartmouth Claude and asked whether it showed signs of AI writing, then incorporated many of the changes Claude suggested. When an editor later asked about AI use, Schnell requested a pause in publication and described the tool as an editorial aid; the committee judged that account understated what the logs showed.
A targeted review, not a disciplinary proceeding
The committee was asked to review Schnell’s AI use against publication rules and Dartmouth policies. Its report says he had published at least 30 writings since taking up his Dartmouth role and listed seven more as forthcoming, but a complete review was not possible in the time available. The panel focused on work for which relevant publisher rules, records of AI use and missing or incomplete disclosures could be assessed.
The committee also reported limits on its fact-finding: Schnell provided his Dartmouth-issued laptop but did not supply requested logs from personal large-language-model accounts, and did not make himself available at the requested time or propose another meeting. The panel said it did not use Pangram or other AI tools to determine whether or how much he used AI. Its conclusions rested on records that included chat logs, emails and publisher policies.
On Dartmouth policy, the committee concluded that Schnell’s conduct could qualify as a violation of the College’s Code of Ethical Conduct, Academic Honor Principle and Geisel School of Medicine professionalism policy. It also stressed that its review did not follow the procedures for assessing formal policy complaints and was not recommending related disciplinary action. Dartmouth’s announced response is Schnell’s departure and Delmont’s interim appointment through June 2028.