Dartmouth Provost Resigns Over AI Use – Inside Higher Ed

Dartmouth University Provost Santiago Schnell resigned Wednesday after a faculty-led review committee concluded that he “used AI to meaningfully alter his writing” and “fell short” of the university’s academic integrity standards.  

“I expect our senior leaders to hold themselves to the highest standards,” Dartmouth President Sian Leah Beilock, who has called for “more AI” at the university, wrote in a memo to students, faculty and staff. “Out of respect for his colleagues and the institution, Santiago has resigned as provost and is leaving the university.”

His resignation comes less than two weeks before he was set to face a no-confidence vote from faculty leaders. 

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. A few weeks later, The Dartmouth followed up with its own investigation into more of Schnell’s published writings, also using Pangram to detect the extent of AI use. According to the student newspaper, Pangram determined that Schnell’s work was 96 percent “AI-written” across the nine professional articles he published this year, including the Post essay and two academic papers about mathematics. 

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.” But that explanation did little to quiet the blowback from students and faculty who argued that he violated the college’s honor code—which requires students to acknowledge if they use generative AI tools to complete their coursework—and should be fired. 

In response to the outcry, Beilock convened an independent advisory committee composed of faculty and alumni to review how Schnell used AI in his writing and determine if he met publishers’ AI-disclosure requirements, if he was honest about his AI use and if that use constituted a violation of Dartmouth policy. 

The Claude Logs

The committee, which sent its eight-page report to Beilock on Tuesday, did not use Pangram or other AI detectors—which experts say can be unreliable—to support the review. Instead, it focused on analyzing articles he wrote for publishers that either barred AI use or required disclosure, and drew on available records of the provost’s communications on the matter or records from his Dartmouth Claude account that indicated that the provost used AI to edit his research and/or writings. 

However, the committee noted that Schnell didn’t fully cooperate with their investigation. 

“Although he did provide his Dartmouth-issued laptop for inspection, the Provost did not, upon request, provide logs of his use of personal LLM accounts that he used for his academic and public writings,” the report reads. “The Provost also failed to make himself available at the time requested or to propose additional times that he could meet to discuss his AI use across his recent writings.”

Nonetheless, the committee was able to look at the conversation he had with the university’s enterprise version of Claude while addressing edits from peer reviewers for a paper later published in the journal ChemSystemsChem last month. It determined that he used the system 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.”  

According to the report, his questions for Claude included, “Do you find any scientific issue with the paper, problem or contradiction that it would prevent me to submit it [sic]?” and “I asked ChatGPT to write a version of the response” to peer reviewers “which you will find enclosed. Please check it. Based on this version would you change anything in your version?” 

The committee also noted the publication’s policies “at the time of the paper’s submission, acceptance, and publication all required disclosure of AI use of this kind,” but the “provost failed to provide any such disclosure.” Although Schnell had previously explained the discrepancy—the publisher only began requiring AI disclosures in 2025 and he was relying on 2025 author guidelines—the committee found that the publisher has required disclosure of AI use since at least 2023.

Claude logs also show that Schnell uploaded the manuscript of a forthcoming article entitled “Academic Freedom and Catholic Mission: The Procedural Overlap Zone in University Governance.” He asked the LLM if there were “many signatures of AI writing in this article?” and, according to the report, “later submitted a revised manuscript that incorporated the kinds of changes Claude suggested.”

After the uproar over his AI use began, Schnell asked the editor to pause publication of the article. Although the publisher doesn’t have a specific AI policy, he disclosed that he “used AI as an editorial tool during the article preparation and revision process to help refine sentences, improve clarity, and the expression of arguments,” according to the report. 

However, the committee said that statement “understates the nature of the provost’s actual AI use according to the Claude logs,” and concluded both papers “went beyond his other public descriptions of his use of AI in writing.” In addition, the committee “felt that the provost’s incomplete (or missing) disclosures about his AI use, and public statements characterizing his AI use, are not consistent with ‘a culture of integrity, honesty, and respect in … scholarship’” laid out in university policy. 

But Schnell’s resignation doesn’t mean that Dartmouth—which calls itself the birthplace of AI—is turning its attention away from the technology. 

“Each of us brings a different perspective on how and when to incorporate AI into our work and our lives. Moreover, keeping Dartmouth at the forefront of AI inevitably means leading through an environment that is fluid and fast-changing,” Beilock said. “That is why, as leaders, it is so critical that we model the transparency and human judgment that we expect from our students and scholars.”

Matthew Delmont, a history professor and associate dean of faculty for international and interdisciplinary studies, will serve as Dartmouth’s interim provost through June 2028.