“They Used AI” Is Not an Argument | American Enterprise Institute
The Chronicle of Higher Education asked Carolyn Rouse, Princeton professor and president of the American Anthropological Association, to respond to the most serious indictment her discipline has faced in a generation.
Vanderbilt and Washington University in St. Louis’s report on the state of humanities scholarship named anthropology “the most extreme case” of politicized decline among the six fields it reviewed. The report found that “reasonable dissent on politically charged topics is routinely suppressed and punished.”
In response, Rouse declared, “They used AI.” These are the first three words of her interview; the president of a major scholarly association reacted to a fair concern and powerful indictment with an accusation of robot authorship and disavowed the report—comparing it to chatbot-written student papers.
Rouse’s retort about AI falls apart when the details of the report are considered, however. Kwame Anthony Appiah, a report author, told the Chronicle that the group used artificial intelligence to survey large bodies of disciplinary text and identify themes within them—a decades old social-science method. However, Rouse converts a description of AI-assisted research into an assertion that the report is machine-written slop. More telling, she concedes the suspicion came first: “Even before we read that, we were all thinking, ‘This reads like they used AI.’”
This rhetorical move is common in academia and professors regularly keep conversation-stoppers on hand—words that end an argument without anyone making one. “Positivist” did this work in the 1990s; “harmful” more recently. “They used AI” is the newest iteration and perhaps the most effective: unfalsifiable, requiring no reading, and shifting the burden to the accused, who must establish their humanity before their evidence can be weighed.
In her Chronicle interview, Rouse’s dismissive reflex was not isolated. When asked about Joseph Henrich—the lone anthropologist on the commission, among the most cited scholars in the discipline’s modern history, and a well-regarded public intellectual—Rouse’s complete answer was a single word: no.
Shown a peer-reviewed survey in which forensic anthropologists split roughly down the middle on whether sex is binary, she replied: “I don’t believe in opinion research. Not to disparage them, but a lot of forensic people, they’re coroners, they’re doing it in a practicing level, where they’re actually asked on forms to determine whether this body is male or female, oftentimes they haven’t had advanced schooling.” Rouse is no position to disqualify the legitimacy of survey methodology and reject a finding.
Rouse went further. When asked whether she regrets the AAA’s 2023 cancellation of an accepted conference panel on biological sex as an analytic category, she answered: “It should never have been accepted … they slacked on the peer review.” Then, Rouse dug deeper by stating: “Things like that are cut all the time in peer review, you never see it. Journalists never see it. I never see it. There’s actually a filter happening all the time. This one just happened to make it into the public space.”
The report’s gravest charge is that anthropology routinely and quietly suppresses reasonable dissent. Rouse did not refute that charge; she described the machinery, called it quality control, and offered its usual invisibility as proof the system works.
I have been writing recently about a different kind of AI skepticism: the one I see in my own classroom. My students at Sarah Lawrence are often deeply wary of these tools—not because AI discredits someone else’s work, but because they worry about what relying on it will do to their own minds and their education. That skepticism is intellectually healthy because it begins with self-scrutiny. Rouse, on the other hand, directs that skepticism outward, dismissing someone else’s argument before engaging with it, which runs directly against traditional scholarly norms of debate and dialogue.
To be fair, Rouse later offers a legitimate methodological critique. She argues that a representative study would require sampling institutions across the political spectrum, examining syllabi, interviewing faculty, and observing classrooms. This is exactly the kind of criticism scholars should make. But rather than making this argument at the onset of her discussion, Rouse tried to delegitimize the report before she professionally evaluated and responsibly discussed the evidence.
This interview captures so much of what is wrong with academia today. Ironically, Rouse concluded her first response by lamenting that “unserious things right now have so much power in the academy.” Yet it was Rouse’s own answers in the Chronicle that came across as unserious and, ultimately, embarrassing for both her discipline and higher education more generally, especially given her role as president of a major scholarly association. Stepping back, colleges and their faculty will not rebuild public confidence by dismissing critics or attacking legitimate research methods. They will rebuild it by engaging criticism honestly, confronting evidence directly, and proving their critics wrong. In this interview, Rouse failed to do so.
