Skip to main content
NOPE
Medium reported severity Internal Action

Dartmouth provost resigns after a faculty-led review of his Claude logs finds undisclosed AI use in published writing

Santiago Schnell, provost of Dartmouth College in Hanover, New Hampshire, resigned and is leaving the college, President Sian Leah Beilock announced on 28 October 2026. The announcement followed the report of an advisory committee she had convened on 25 September after Semafor and the student newspaper The Dartmouth reported that the AI detector Pangram flagged his 2026 op-eds and articles as AI-written. The committee did not use Pangram. Drawing on his Dartmouth Claude enterprise logs and his emails, it found that he used AI to meaningfully alter his writing in violation of at least one publication's AI disclosure requirement, failed to disclose ChatGPT use in a Washington Post column against the newspaper's policy, and was not forthcoming about his AI use. The committee said it was not recommending disciplinary action, and the president said the resignation was his decision. Inside Higher Ed reports that he was due to face a faculty no-confidence vote. He has said AI did not substitute for his own research, judgment or authorship.

AI system
Claude and ChatGPT
Occurred
28 Oct 2026
Reported
28 October 2026
Event location
United States
What the AI did
Relation unknown
Reported harm
Professional HarmReputational Harm
Whose AI use
Their own AI use
Setting
Work · Education
Evidence
AI involvement reported · Causal attribution supported · 6 sources
7 claims: 4 documented, 3 reported. 3 open questions
People reported harmed
1 person

AI system as recorded: Anthropic's Claude (Dartmouth's enterprise deployment), whose logs the committee reviewed, and OpenAI's ChatGPT, which the provost said he used on a Washington Post column; he also used personal LLM accounts whose logs he did not provide

What Happened

Background. In August 2026 Semafor reported that the AI detector Pangram flagged a Washington Post column by Dartmouth's provost as AI-generated. On 21 September The Dartmouth reported that Pangram gave nine of his 2026 publications a median AI-written score of 96%. On 25 September the president convened an independent, faculty-led advisory committee to review his AI use.

The review. The committee's report of 27 October says it did not use Pangram or other AI tools. It reviewed his Dartmouth Claude enterprise logs, his emails and his correspondence with publishers, and says he did not provide logs of the personal LLM accounts he used for his writing. It found that he used Claude to substantially edit a chemistry paper, troubleshoot code and generate figures after peer review, without the disclosure the journal required; that his admitted ChatGPT use on the Washington Post column clearly violated the newspaper's policy; and that his description of AI use on a forthcoming article understated what the Claude logs showed. For a third paper it could not conclude there was a violation. It found his conduct fell short of Dartmouth's Code of Ethical Conduct, the Academic Honor Principle and a Geisel School of Medicine professionalism policy, while stating that it was not proceeding under the complaint procedures for those policies or recommending disciplinary action.

The resignation. On 28 October the president wrote that he had resigned as provost and was leaving the university, that this was his decision, and appointed an interim provost. Inside Higher Ed reports that the resignation came less than two weeks before a planned faculty no-confidence vote.

His account. He told The Dartmouth that he used AI to make writing more efficient but that it did not substitute for his intellectual judgment or authorship.

Reported harm

The provost lost his post and is leaving Dartmouth after a college committee found undisclosed AI use in his published writing; the president says the resignation was his decision.

Outcome

Resolved

The provost resigned and is leaving Dartmouth; an interim provost was appointed through June 2028. The advisory committee stated it was not recommending any disciplinary action, and the president wrote that the resignation was his decision.

What remains unknown

  • What his personal LLM accounts were used for: he did not provide their logs to the committee.
  • Whether he would have been removed had he not resigned: the committee recommended no disciplinary action and the no-confidence vote did not take place.
  • How much of the published text the AI tools produced: the committee relied on the logs it could read, and the Pangram scores reported by outlets are detector estimates.

What the evidence supports

AI involvement: reported. Consequential response: the harm is the college's response to the provost's own use of AI in published writing. The committee report quotes his prompts to Dartmouth's enterprise Claude and finds he used it to substantially edit a manuscript, troubleshoot code and generate figures without the disclosure the journal required, and the president announced his resignation the day after the report. No source says an AI output was itself the cause of an adverse decision about him, so no relation is recorded; the Pangram detector scores that started the controversy were not used by the committee.

7 claims: 4 documented, 3 reported. What the statuses mean

Reported Semafor and The Dartmouth reported in August and September 2026 that the AI detector Pangram flagged the provost's 2026 writing, including a Washington Post column, as AI-written.

Causal attribution. Background to the review; the detector scores are the outlets' tests and were not used by the committee.

  • thedartmouth.com(opens in new tab) supports · English
    'An investigation by The Dartmouth of academic articles and opinion pieces published by Provost Santiago Schnell in 2026 found a median “AI-written” text percentage of 96% using artificial intelligence detector Pangram.'; 'identified a column by Schnell as “100% AI-written.”'
  • insidehighered.com(opens in new tab) supports · English
    'The controversy surrounding Schnell’s AI use started in August when Semafor published an investigation showing that the AI detector Pangram flagged an editorial he wrote for The Washington Post as at least 80 percent AI generated.'
Documented On 25 September 2026 the president convened an independent advisory committee to review the provost's AI use; the committee did not use Pangram or other AI tools in its findings.

Causal attribution. Established by the committee's own report.

  • president.dartmouth.edu(opens in new tab) supports · English
    'On September 25, 2026, President Beilock convened the Advisory Committee to review Provost Schnell’s AI use in his academic publications and public writings.'; 'The Committee did not use Pangram or other AI tools when determining whether and to what extent Provost Schnell used AI'
Documented The committee found from his Dartmouth Claude logs that he used Claude to substantially edit a research manuscript, troubleshoot code and generate figures after peer review, without the AI disclosure the journal required, and that he did not provide logs of his personal LLM accounts.

Causal attribution. The committee's finding, documented by its report; the underlying logs were not published in full.

  • president.dartmouth.edu(opens in new tab) supports · English
    'The Committee has determined that the Provost used Dartmouth’s enterprise AI tool, Claude, logs of which the Committee reviewed, to substantially edit his research manuscript, troubleshoot errors in computer code, and generate figures for his publication following the receipt of major revisions from peer reviewers.'; 'the Provost did not, upon request, provide logs of his use of personal LLM accounts that he used for his academic and public writings.'
Documented The committee concluded that he used AI to meaningfully alter his writing in violation of at least one publication's requirement, failed to disclose AI use against a national newspaper's policy, was not forthcoming, and fell short of Dartmouth standards, while not recommending disciplinary action.

Causal attribution. The committee's conclusions as stated in its report.

  • president.dartmouth.edu(opens in new tab) supports · English
    'this Advisory Committee finds that the Provost used AI to meaningfully alter his writing in violation of at least one publication’s AI use and/or disclosure requirement, failed to disclose his use of AI in clear violation of a national newspaper’s stated policy, was not forthcoming about his AI use'; 'nor recommending any related disciplinary action.'
Documented He resigned as provost and is leaving the university; the president says the decision was his.

Causal attribution. Established by the president's own statement; the link to the review is the statement's own sequence.

  • president.dartmouth.edu(opens in new tab) supports · English
    'Out of respect for his colleagues and the institution, Santiago has resigned as provost and is leaving the university. This was his decision.'
Reported The resignation came less than two weeks before he was due to face a faculty no-confidence vote.

Causal attribution. Single outlet; the vote is not described in the other sources read.

Reported He has said that he used AI to assist his writing but that it did not substitute for his research, judgment or authorship.

Causal attribution. His own account, relayed from one interview with the student newspaper; counterevidence to the scale of AI use.

  • thedartmouth.com(opens in new tab) supports · English
    'Schnell said he used artificial intelligence for “making writing processes more efficient,” but that AI was not a “substitute” for his own “intellectual judgment or authorship.”'
  • insidehighered.com(opens in new tab) supports · English
    'Although Schnell admitted to using AI to assist in writing the articles, he told the student newspaper that it didn’t “substitute” for his “research, intellectual judgment or authorship.”'

Sources

6 sources inspected. Sources that repeat one account do not corroborate each other.

How the sources were read, and where the events happened

Read in English on 2026-10-29 (HTTP 200). The president's statement announcing the resignation and the report. Applies to s1.

Read in English on 2026-10-29 (PDF, HTTP 200, all eight pages). The committee's own findings; it quotes the provost's prompts to Claude and his emails. Applies to s2.

Read in English on 2026-10-29 (HTTP 200). Student newspaper report on the resignation and the committee report. Applies to s3.

Read in English on 2026-10-29 (HTTP 200). Relies on the president's memo, the committee report and earlier reporting. Applies to s4.

Read in English on 2026-10-29 (HTTP 200). Local newspaper report relying on the president's email and the committee report. Applies to s5.

Read in English on 2026-10-29 (HTTP 200). The student newspaper's own Pangram test of the provost's writing and his response. Applies to s6.

Event countries: United States. Affected-person countries: Unknown. Court countries: Unknown.

The review and resignation took place at Dartmouth College in Hanover, New Hampshire (Valley News dateline HANOVER, president's statement). His own country of residence is not used.

Reviewed for publication 2026-10-29: Published as a consequential response: the college's own committee report documents his Claude use and undisclosed AI use in published writing, and the president's statement records his resignation the next day. His account that AI did not replace his judgment, and the committee's statement that it recommended no discipline, are recorded.

People described

The provost of Dartmouth College, a senior university administrator and public figure, who resigned after the college's review of his AI use

People reported harmed in this case

1 person

1 AI participant · 0 other people harmed

One person: the provost, who used the AI tools himself and resigned after the review (president's statement, committee report). Co-authors and editors are not counted.

Counted once within this case. The same person may appear in other cases. This count does not establish AI causation.

Tags

#work#higher-education#resignation#undisclosed-ai-use#institutional-review

Cite this case

Compiled per our published methodology: verification statuses, sourcing standards, and corrections process.

APA

NOPE. (2026). Dartmouth provost resigns after a faculty-led review of his Claude logs finds undisclosed AI use in published writing. AI incidents. https://nope.net/incidents/2026-dartmouth-provost-resigns-after-college-review-finds-undisclosed-ai-use-in-published-writing

BibTeX

@misc{2026_dartmouth_provost_resigns_after_college_review_finds_undisclosed_ai_use_in_published_writing,
  title = {Dartmouth provost resigns after a faculty-led review of his Claude logs finds undisclosed AI use in published writing},
  author = {NOPE},
  year = {2026},
  howpublished = {AI incidents},
  url = {https://nope.net/incidents/2026-dartmouth-provost-resigns-after-college-review-finds-undisclosed-ai-use-in-published-writing}
}

Related cases

Medium Unidentified image tool

Los Banos, California: a substitute teacher says a sexualised AI image of him made from an old selfie circulated among middle-school students on TikTok and Snapchat and was used by a student account impersonating him; one student received in-school suspension and he stopped taking assignments at the middle school

WIRED (Caroline Haskins, 24 August 2026) reports the account of a substitute teacher in Los Banos, California, who spoke to WIRED by name (not repeated here). On 12 March 2026 a co-worker at his other job told him an AI photo was circulating on social media; the image reused a mirror selfie he had taken at a job in college but showed him in lingerie, posing suggestively. He traced an account sharing it on TikTok to a student at the middle school where he regularly substituted and reported it to the assistant principal; that evening he found a second account, created by a different student, impersonating him under a misspelled version of his name with the image as its profile picture. Over the following weeks he dealt with administrators, parents and the police; one student who admitted creating the image received an in-school suspension. He says the image was also shared on Snapchat, kept spreading among students and drew crude sexual comments on his TikTok account. He told administrators he would not take assignments at the middle school until the matter was addressed; at the local high school the same image was AirDropped to him during a class. WIRED says the district did not respond to its request for comment. He said he does not feel safe returning to the district. The Los Banos Enterprise (25 August 2026) summarised the WIRED account; no independent local reporting was found.

Medium Jewish Onliner (suspected)

Yale Law School scholar placed on leave and barred from campus after article on an AI-empowered news site linked the scholar to a sanctioned group

In March 2025 Yale placed a Yale Law School scholar, who was deputy director of a Yale Law School project, on administrative leave and barred the scholar from campus. The New York Times reports the decision came three days after a news site described as powered at least in part by artificial intelligence published a story on the scholar's connections to Samidoun, a group on a US sanctions list. Inside Higher Ed and Middle East Eye identify the site as Jewish Onliner. The scholar's lawyer says Yale's general counsel named the article as the trigger of the investigation. The scholar told the Times of not being a member of any organization that would violate US law, and the lawyer says the scholar is not a member of Samidoun. A Yale Law School representative told The National that placing an employee on temporary administrative leave while a review is conducted is the appropriate process and that the scholar's short-term position was due to expire the following month. The site says humans fact-check and that AI tools play a significant role in its work. The inspected sources do not show that AI produced the article.

Medium Unidentified voice-cloning tool

Baltimore County, Maryland: police say a high school's athletic director framed the principal as racist with an AI-cloned voice recording

Police in Baltimore County, Maryland, said in April 2024 that a high school principal had been framed as racist by a fake recording of his voice containing racist and antisemitic comments, which appeared in teachers' email inboxes and spread on social media. Court records cited by the Associated Press say an expert consulted by detectives found traces of AI-generated content with human editing; a second expert found spliced recordings and told AP he had not confirmed the clip was fully AI-generated. The principal went on leave while police guarded his house. The school's athletic director, investigated for using AI to make the recording in retaliation, entered an Alford plea, maintaining his innocence, and was sentenced in 2025 to four months for disturbing school operations, CBS Baltimore reports.

Medium ChatGPT

Saginaw, Michigan: a credit union CEO used ChatGPT to put herself and family members in 'Lake America' sweatshirts in a photo taken before a trip to Halifax; after her sister reposted it publicly without its AI label she says she received death threats and the family came home early, and on 23 September 2026 the credit union said she was no longer employed

The chief executive of Family First Credit Union in Saginaw, Michigan, told WJRT (ABC12) that before boarding a flight to Halifax, Nova Scotia, on 11 September 2026 she and family members took a photo and put it through ChatGPT to show them wearing 'Lake America' sweatshirts, a joke about the US president's order renaming Lake Ontario. She posted it to her private Facebook page with a marker saying it contained AI content; her sister reposted it publicly without the marker while they were in Halifax, and it spread in Canada as if the family had worn the shirts. The backlash concerned the image's political message; she said she would understand the anger of anyone in Halifax who thought the family had walked in wearing those sweatshirts. She said she was getting death threats and the family returned early; by 16 September she was back in the US. She called the post poor judgment, said she would not use AI again and that AI 'can make people think something's real that's not'. On 23 September the credit union said she was no longer an employee, effective immediately; it has not said whether she resigned or was dismissed, or why.

If you or someone you know is struggling, free and confidential support is available. Find a helpline near you at Signpost.

Back to all incidents