College Leaders Are Using A.I. Students Call It Hypocrisy.
By Gammatek ISPL Last updated: September 2026 | 14 min read
Why This Matters Right Now
If you run any organization where policy has to apply evenly to be credible — a company, a school, a compliance program — a story breaking out of American universities this month should get your attention. Several college and university leaders who built their public reputations warning about the risks of student AI use have been caught, sometimes by their own students, using AI themselves without disclosing it. The fallout isn't just an academic controversy. It's a live case study in what happens when the people writing the rules don't follow them — and it's playing out at an institutional scale that any organization drafting its own AI policy should be watching closely.
The Dartmouth Case That Started It
The clearest example comes from Dartmouth College. Provost Santiago Schnell published a Washington Post op-ed arguing universities should move past policing student AI use and instead redesign how they assess student work. Student reporters at the campus paper, The Dartmouth, later analyzed his individually authored 2026 academic and opinion writing and found a median AI-generated-text estimate of 96 percent across those pieces — a finding that, once published, set off what one outlet described as a campus firestorm among faculty and students who felt his after-the-fact, piecemeal disclosures amounted to hypocrisy.
The timing made it worse. Dartmouth's president, Sian Beilock, had recently written a piece for The Atlantic titled "Why I Want More AI at Dartmouth," arguing that students and faculty should actively practice using the technology to prepare for a future where it becomes even more central to how people work. Dartmouth's own student AI policy requires disclosure whenever a professor permits AI use in coursework — a standard that, students pointed out, the provost's own writing hadn't obviously met.
A Dartmouth history professor summarized the underlying tension bluntly, noting the university's identity as a liberal arts institution sits uneasily with senior leaders modeling exactly the behavior the school asks students not to do. A university alum was more direct, describing what he saw as "a tremendous hypocrisy" in the gap between how campus leaders talk about ethical AI use and how they apparently practice it themselves.
It's Not Just Dartmouth
Once the Dartmouth story broke, similar tensions surfaced elsewhere, suggesting this isn't an isolated incident but a structural pattern across higher education.
At the University of North Carolina system, leadership sent an official letter — as part of a broader "Compact 2.0" response — asserting that universities have an obligation to verify students have genuinely mastered what their degree represents, and calling on instructors to set clear guidance requiring students to disclose AI use. What the letter didn't mention: AI had reportedly been used to help draft the letter itself. A self-described "AI resistor" at the university called the letter hypocritical on exactly those grounds.
Separately, a University of California, Berkeley math professor who wrote a widely shared column criticizing student preparedness later acknowledged using AI to help edit that same column, according to campus reporting.
At Harvard, Dean David Deming argued in a message to faculty and students that policing AI use was actively damaging faculty-student trust, urging "acceptance, or even encouragement" of AI for writing-intensive coursework — a position that, paired with the Dartmouth revelations, added fuel to a broader debate about whether campus leadership is asking more of students than of itself.
Institution | Leadership's public stance | What surfaced | Student disclosure requirement |
Dartmouth | President urged wider AI adoption; provost wrote critically on AI policy | Provost's own writing estimated at 96% AI-generated, undisclosed | Required when professor permits use |
UNC System | Letter demanded instructors enforce student AI disclosure | Letter itself apparently drafted with AI, undisclosed | Required under new "Compact 2.0" guidance |
UC Berkeley | Professor publicly criticized student preparedness | Acknowledged using AI to edit the column after publication | Standard academic integrity policy |
Harvard | Dean urged broader classroom acceptance of AI | No individual hypocrisy case, but framing seen as inconsistent with stricter student-facing rules elsewhere | Varies by course |
Why This Keeps Happening
A few structural reasons explain why this pattern is emerging now, rather than being a series of unconnected incidents:
Disclosure rules were built for students, not administrators. Most university AI policies were drafted with a classroom scenario in mind — a student submitting an assignment. Nobody wrote an equivalent disclosure requirement for a provost's op-ed or a system-wide policy letter, which left senior leaders operating in a genuine gray zone even before questions of good faith come into it.
AI detection tools focus downward, not upward. Universities invested heavily in tools to flag AI-generated student writing. Almost none built equivalent scrutiny into how administrators produce public-facing communications — so the asymmetry in enforcement isn't just cultural, it's built into the infrastructure.
The incentive to use AI quietly is the same at every level. A provost drafting an op-ed under deadline pressure faces the identical incentive a student facing a paper deadline does. Nothing about seniority removes that pressure — but the accountability structures built around it assume it does.
Students are unusually well-positioned to catch it. Ironically, the same generation of students being told to disclose AI use are also the ones most fluent at detecting AI-generated patterns in writing — which is exactly how the Dartmouth case was uncovered in the first place, by student reporters analyzing the provost's own published work.
The Deeper Debate This Reveals
Underneath the individual controversies is a genuine, unresolved argument about what AI policy in education should even look like. One school of thought, represented by leaders like Beilock and Deming, holds that resistance is largely futile — AI isn't going away, so the more productive path is teaching people, students and leaders alike, to use it well and transparently, rather than pretending prohibition works. Critics counter that this framing conveniently excuses inconsistent enforcement, and that "practice makes acceptance" arguments tend to arrive suspiciously often from the same people whose own AI use has just been scrutinized.
There's a third, less discussed layer to this: students aren't uniformly pro-AI either. At several 2026 commencement ceremonies, students booed speakers who used their remarks to promote AI enthusiastically — incidents at the University of Central Florida and Marquette University drew particular attention. A student organizer with a campus chapter of an AI-caution advocacy group argued students were right to call out what she saw as hypocrisy from a "pro-AI crowd" that rarely faces the same scrutiny it applies to students.
An Implementation Consideration for Any Organization Writing AI Policy
This isn't just a university problem, and it's worth pausing on why. Any organization drafting AI use policy — a company, a school, a regulated industry — runs into the same structural gap universities just got publicly caught in: policies written for the people being managed rarely get written with equal force for the people doing the managing.
A genuinely credible AI policy needs a few things university policies mostly lack right now:
Disclosure requirements that apply at every level, not just to the people with the least power to push back if caught.
A defined process for how leadership's own AI-assisted work gets reviewed — not just student submissions.
Public-facing communications treated under the same rules as internal ones — an op-ed or an official letter is exactly the kind of high-visibility, high-stakes writing where undisclosed AI use does the most reputational damage if discovered later, as Dartmouth's case shows.
Organizations that build AI governance policy only around monitoring the people at the bottom of an org chart are building the exact structural weakness that just became a national story in higher education.
What Comes Next
None of this means AI has no place in higher education leadership — Beilock and Deming's underlying argument, that pretending prohibition works is naive, isn't unreasonable on its face. The actual failure in these cases wasn't using AI. It was asking for a standard of disclosure from students that leadership didn't hold itself to, and getting caught not holding itself to it. That's a governance failure, not a technology failure — and it's exactly the kind of gap that shows up wherever policy exists on paper but enforcement only flows in one direction.
For any organization watching this play out and wondering whether its own AI policy has the same blind spot, the Dartmouth story is a useful, uncomfortable mirror: it's worth checking now, before students, employees, or auditors find the gap for you.
How This Connects to Governance Beyond Campus
The same asymmetry showing up in higher education — rules written for one group, quietly unenforced for another — is a familiar risk in any regulated organization managing AI adoption, from manufacturing compliance teams to plant safety audits. A policy that can't survive equal scrutiny at every level isn't really a policy; it's a liability waiting to be discovered.
[See how Gammatek helps organizations build AI and compliance policies that hold up under audit, at every level →https://www.gammateksolutions.com/post/it-s-all-fun-and-games-until-you-give-ai-your-credit-card
Comments