Dartmouth College announced an investigation into its own provost, Santiago Schnell, after the college’s student newspaper found that some of his published work appeared to have been written by artificial intelligence. The provost who helped oversee an institution that penalizes students for undisclosed AI use had, according to reporting, been doing precisely that himself.
Institutional Disclosure Rules That Apply Differently at the Top
The Geek.hr Team, which follows technology and science developments for Croatian readers, sees the structural problem in the Dartmouth case as direct and difficult to dismiss. The same administrators who have positioned AI disclosure as an academic integrity requirement for students have not consistently held themselves to that standard.
Croatian technology coverage by GEEK.HR has tracked how this same institutional tug-of-war over who must disclose AI use plays out well beyond American campuses, documenting the disclosure-and-integrity tensions that universities in multiple countries are still working through. The Dartmouth case makes that dynamic unusually visible.
“The pattern in the United States maps onto a wider question that technology observers in other markets have been watching closely: when institutions write the rules, do those rules bind the rule-writers?”
The George Washington University working group finding from August 2026, that many students there had expressed doubts about whether faculty were using AI ethically and transparently, suggests the Dartmouth provost’s case is not an isolated grievance but a specific, documented instance of a broadly felt suspicion.
The Schnell Investigation, From Op-Ed to Apology
The sequence of events is precise and carries its own irony at nearly every step. Semafor’s investigation in August 2026 found that a commentary Schnell published in The Washington Post on Aug. 9 of that year, an opinion piece specifically about AI cheating in schools, had itself been written by AI. An administrator writing about the harms of student AI misuse, without disclosing that he had used AI to write the piece, was the core of the problem.
Boston.com reported that The Dartmouth’s follow-up investigation in September 2026 found extensive AI use across Schnell’s published work that year. When confronted, Schnell told the student paper he had used AI “for language refinement, copyediting, improving clarity and organization and making writing processes more efficient.” The Dartmouth then ran that explanation through an AI detector, which flagged the statement itself as apparently AI-written.
On Sept. 25, Schnell published a letter in The Dartmouth apologizing for not disclosing his AI use consistently in public writing. “Suffice it to say, I am more aware than ever of the need for robust intellectual discussions about the appropriate use of AI in the higher education landscape,” he wrote. Dartmouth President Sian Beilock, who had recently called for expanded AI use at the university, announced the investigation and acknowledged the difficulty of leading on new technology. “There will be missteps,” Beilock wrote, “and that’s part of how an institution learns.”
Dartmouth senior Eli Moyse had a more direct response. “It just goes against all the values of our institution and the concept of academia,” he said. He added that Schnell “should resign immediately or get fired,” calling the conduct “plagiarism, even if it’s written by a robot.”
Student Suspicion and Faculty Use Are Both Measurable
The GWU finding did not emerge from a single complaint. The working group there reported in August 2026 that student doubt about faculty AI ethics was widespread, not anecdotal.
Faculty behavior data adds texture to that concern. A survey conducted last fall by Elon University and the American Association of Colleges and Universities, covering 1,057 faculty members nationwide, found that roughly a quarter were not using generative AI tools at all. Close to half said they turned to AI at least somewhat when creating or revising assignments. Just 6 percent reported using AI a lot when preparing their own research papers, and nearly 9 in 10 said they did not use AI for grading.
Meanwhile, more than half of U.S. college students report using AI at least once a week. One in five say they use it every day. The asymmetry between how often students use these tools and how closely administrators monitor that use, while exempting institutional communications, is what student critics describe as the core inequity.
Berkeley and Stanford Add to the Pattern
The Dartmouth case is not alone. The Daily Californian, Berkeley’s student newspaper, ran dozens of campus communications through AI detectors and found that some messages from top administrators appeared to have been written or altered by AI. Berkeley spokesperson Dan Mogulof said administrators were using AI within certain guidelines, “so long as the final product accurately and authentically captures and expresses the author’s thoughts, perspectives, tone and objectives.” That standard, notably, is one institutions have rarely applied symmetrically to student work.
Stanford University presented a separate category of institutional AI misuse. The university used AI to remove a Hispanic man from a photograph, replacing him with a Black woman, and to alter the appearances of two other students in the same image. After the changes were discovered, Stanford acknowledged a policy violation.
Policy Attempts Exist, but a Common Standard Does Not
The University of Southern California has moved toward establishing disclosure guidelines, offering an example of what a formal policy attempt looks like. But such efforts remain scattered and inconsistent across institutions.
Marc Watkins, who founded the Mississippi AI Institute for Teachers at the University of Mississippi, put the problem plainly. “The big thing is there’s no social contract right now,” he said. “We don’t have these sorts of priors to say what’s appropriate and what’s not.”
That absence of shared prior agreement is what makes the Dartmouth investigation something more than a personnel matter. The case is ongoing, its outcome still undetermined. What it has already established is that institutions have set disclosure expectations for students while leaving their own leadership without equivalent obligations. Watkins’ framing holds: no community agreement exists yet on what counts as appropriate, and until one does, the gap between what administrators require and what they practice will keep producing exactly these confrontations.
The post Dartmouth Provost’s AI-Written Work Puts Campus Double Standard Under Scrutiny appeared first on Daily Emerald.