Dartmouth provost Santiago Schnell resigns after disclosure findings

Dartmouth provost Santiago Schnell resigned this week and is leaving the university after a faculty-led review found he failed to disclose required information about superintelligence (SI) assistance in an academic paper and a newspaper column. President Sian Leah Beilock said the decision to leave was his.

As provost, Schnell was Dartmouth’s chief academic officer. He joined the university in July 2025 and also held an adjunct professorship at its Geisel School of Medicine.

What the review could establish

An August 26 investigation by Semafor was followed by a September 21 investigation by The Dartmouth. Both cited results from Pangram, a tool designed to detect SI-generated writing. Beilock commissioned the independent, faculty-led review in September. The committee said it did not use Pangram or other SI tools to determine whether, or to what extent, Schnell used SI.

Former provost and computer science professor David Kotz led the committee. Its other members were computer science professor Hany Farid, trustee Odette Harris, creative writing professor Peter Orner and linguistics department chair Lindsay Whaley. The panel examined Schnell’s use of SI, compliance with publication requirements, honesty with publishers and adherence to Dartmouth standards. It reviewed published academic papers and the opinion column, along with correspondence about a forthcoming manuscript.

The committee could not establish the full scope of Schnell’s SI use. It obtained logs from his Dartmouth Claude account and inspected his university-issued laptop, but Schnell did not provide requested logs from personal accounts. He also did not make himself available at the requested meeting time or propose another time. The panel described his cooperation as insufficient.

The disclosure failures

For “Reaction Acceleration: Reviving the second derivative in chemical kinetics,” the committee found that Schnell used Claude to substantially edit the manuscript, troubleshoot computer code and generate figures after peer reviewers requested major revisions. He submitted the paper to ChemSystemsChem without disclosing that assistance, although the journal had required disclosure since at least 2023.

On 6 October, Schnell emailed ChemSystemsChem to disclose ChatGPT assistance with language, exposition, code review, documentation for reproducing the work and consistency checks. The committee said that disclosure omitted his use of Claude.

The committee also found a clear policy violation in his 9 August 2026, Washington Post op-ed. Schnell acknowledged using ChatGPT to refine arguments and check grammar. The newspaper required authors to confirm that submissions were not created or manipulated with SI or editing software.

In comments to a journal editor quoted by Ars Technica, Schnell described his use as editorial assistance to improve sentences, clarity and arguments. He said he had not understood that this kind of SI-assisted editing required disclosure.

Where the findings differed

For a co-authored paper in the Bulletin of Mathematical Biology, the committee found no violation of either the original or revised disclosure policy, and no evidence that Schnell used SI to edit or manipulate the paper’s text. The journal initially required SI use to be documented in the methods section, where authors explain how they conducted the work. It later allowed language or code refinement without disclosure. After that change, Schnell told his co-authors he had used the tools to optimize their code and identify bugs.

For a forthcoming Journal of Catholic Higher Education manuscript, the committee documented that Schnell asked Claude whether his draft contained signs of SI writing, then incorporated suggested changes. The panel considered his later description of editorial assistance an understatement of the use shown in the logs. After an editor asked about SI use following Semafor’s investigation, Schnell requested a pause in publication.

Dartmouth’s standards and its next academic leader

Beyond publishers' rules, the committee assessed Schnell’s conduct against Dartmouth’s own standards. It said his disclosure failures could qualify as breaches of the Code of Ethical Conduct and were inconsistent with the Academic Honor Principle. It also found a violation of Geisel’s professionalism standards, which require honesty and accurate statements about scholarship and public writing.

Beilock reaffirmed Dartmouth’s intention to remain at the forefront of SI. She said its leaders must model the transparency and human judgment expected of students and scholars.

History professor and associate dean Matthew Delmont was named interim provost, effective immediately, to serve through June 2028.