The Dartmouth provost, Santiago Schnell, admitted using AI to assist in writing published opinion pieces, including one about AI-enabled student cheating, sparking student outrage. Analyses by Semafor and The Dartmouth found a Washington Post column flagged as 100% AI-written and nine post-ChatGPT pieces with a median 96% AI score on the Pangram detector. Schnell says he drafted and researched the pieces and used AI only for language and clarity, while students and campus editors say undisclosed use violates academic-honor norms and have called for his dismissal.
Dartmouth Provost Admits Using AI to Draft Op-Eds About AI Cheating; Students Demand Dismissal
Photo Credit: iStock
Provost Santiago Schnell of Dartmouth College has come under sharp criticism after acknowledging he used artificial intelligence to assist in writing opinion pieces — including an op-ed about students using AI to cheat. The revelation prompted student outrage and calls for his removal, with critics arguing the undisclosed use of AI undermines his role as an academic integrity steward.
Reporting by Semafor and follow-up checks by The Dartmouth and Futurism found striking detector results. A Washington Post column by Schnell registered as 100% AI-written when analyzed with the Pangram detector, according to Semafor. The Dartmouth compared Schnell's work from before and after the launch of ChatGPT (November 2022): earlier pieces scored 100% human-written, while nine pieces published after ChatGPT had a median score of 96% AI-written.
Schnell responded that his use of chatbots was limited and that he remained the author of his work. He told reporters that he "develop[s] ideas and arguments, conduct[s] the research, evaluate[s] the evidence and sources, determine[s] the structure, prepare[s] drafts, and decide[s] what appears in the final text." He described AI as an "assistive tool," not a substitute for his judgment or authorship.
Schnell: "I develop ideas and arguments, conduct the research, evaluate the evidence and sources, determine the structure, prepare drafts, and decide what appears in the final text. I regard these technologies as assistive tools; they are no substitute for my research, intellectual judgment, or authorship."
After The Dartmouth ran Schnell's statement through the same Pangram detector, that statement itself registered as 100% AI-written. Schnell later clarified he first drafted the pieces himself and then "used an AI tool to assist with language and clarity," adding, "I am fully responsible for any text under my signature."
Students and campus opinion leaders reacted strongly. Eli Moyse, opinion editor at The Dartmouth, penned an editorial titled "Fire Provost Santiago Schnell," arguing that undisclosed reliance on AI for published work—especially when warning students about AI-assisted cheating—violates the college's Academic Honor Principle and erodes trust.
Eli Moyse: "If students at Dartmouth used AI to write an entire paper without acknowledgement and then submitted it, they would be in clear violation of the Academic Honor Principle. Publishing similarly produced work without disclosure is an insult to our community and readers."
The episode feeds a wider debate over how AI is reshaping academic assessment, integrity policies, and public trust in scholarship. Institutions nationwide are experimenting with new safeguards—from personalized oral exams to revised honor codes—while journalists and researchers race to understand what AI-assisted authorship means for transparency and accountability.
What's next: The controversy may prompt internal reviews at Dartmouth and renewed discussions across higher education about disclosure requirements for AI use, authorship standards, and how to detect and evaluate AI-generated text.
Help us improve.
