Santiago Schnell arrived at Dartmouth’s AI reckoning as one of its most visible advocates. In August 2026, the provost published a Washington Post opinion essay warning that universities confronting AI cheating faced a deeper problem. He later told Semafor he had used chatbots only “to refine a few arguments” and “check for grammatical errors.”¹

But Dartmouth’s review of his work painted a broader picture. The faculty advisory committee examined logs from the university’s Claude enterprise account and found Schnell using the tool to revise research manuscripts, troubleshoot code and edit or generate figures for a paper submitted to ChemSystemsChem. The journal required disclosure of that kind of AI assistance; Schnell did not provide it at submission or publication.²

The committee also concluded that Schnell had failed to disclose AI use in the Post essay, despite the newspaper’s rules. Its inquiry was complicated by what it called insufficient cooperation: Schnell handed over his Dartmouth laptop but did not provide personal LLM-account logs connected to the work, and he declined to meet with the committee.¹

Schnell’s own account of AI use rests on a different distinction. In a later paper on AI detection, he argued that human and machine authorship cannot be reduced to a simple binary: AI can help generate drafts, review arguments and improve language, while the human author retains responsibility for the final work. He rejected calling such tools co-authors, writing that “contribution and authorship are different roles.”¹

That argument did not resolve the disclosure issue. This week, Schnell resigned after the committee determined that he had violated publication requirements and Dartmouth’s ethical standards. President Sian Leah Beilock said the departure was his decision and named historian Matthew Delmont interim provost through June 2028.²

Beilock’s message was not a retreat from AI. Dartmouth, birthplace of the 1956 conference that coined the field’s name, still intends to lead. But she framed the episode as a test of leadership: in a “fluid and fast-changing” environment, officials must model “the transparency and human judgment” expected from students and scholars.²