Zvezdelina Stankova published an op-ed on 15 August arguing that Berkeley admits students who cannot do middle school mathematics. Within days, the argument had been overtaken by a question about how the op-ed itself was written.

Zvezdelina Stankova, a teaching professor of mathematics at UC Berkeley, made her case in the San Francisco Standard under the headline “I teach calculus at Berkeley. Some of my students can’t do middle school math”.

The piece travelled quickly, picked up by Fox News and Townhall, and became a set-piece in the long fight over the University of California’s test-blind admissions policy.

Then Pangram, the detector Substack uses to flag machine-written posts, was pointed at it and returned a reading of roughly 33% AI-generated or AI-assisted. A post about the result by Chris Hoofnagle, a professor at Berkeley Law, drew close to three million views on X, and the story stopped being about calculus.

Stankova has not denied using the technology. She told the Daily Californian that she used AI to help edit the piece, and that the article represented “several hundred person-hours of intensive human work and deliberation, of which about 80 hours are my own”.

That defence is doing more work than it first appears. Editing assistance is, on most campus policies, entirely permissible, and the distinction between a tool that fixes your sentences and one that writes them is exactly the distinction universities have spent three years failing to codify.

Not everyone at Berkeley was persuaded. Hannes Bajohr, who teaches German at the university and writes on machine authorship, told the paper the episode “seems like deception”, or at least something dishonest, absent a disclosure at the point of publication.

The underlying op-ed makes claims that deserve their own scrutiny, and they are not trivial ones. Stankova reports that before 2020, when tests were still required, 71% of her Calculus I students were ready or nearly ready for the course, and that by 2023 the figure was 26%, with the most common diagnostic score being zero.

She also cites admissions disparities between Bay Area schools and notes that five Nobel laureates, among them Jennifer Doudna, have signed open letters calling for standardised tests to return. Those figures come from her own classroom diagnostics and from campaign material, and the university has not published a rebuttal of either.

What makes the episode awkward rather than merely embarrassing is the subject matter. An argument about intellectual rigour, written partly with a tool that students are disciplined for using, is the kind of thing that would be too neat for fiction.

It also lands in a semester when Berkeley’s own computer science faculty have reported rising failure rates alongside heavier AI use in coursework, which suggests the campus is running two conversations about the same technology and has not connected them.

The detection question is not settled either. Pangram is among the more credible tools in a thin field, but independent researchers have argued its false-positive rate is understated, and a 33% score is a probability estimate rather than a confession.

Detectors have been asked to arbitrate authorship in stranger places, including a project that ran the Pope’s writings through them for certification. What none of them can currently do is distinguish a lightly edited human draft from a heavily prompted machine one, which is the case that actually matters.

Faculty elsewhere have taken the enforcement problem into their own hands. A Brown University professor moved his assessments back into supervised in-person exams and produced, in effect, a controlled measurement of how much of his students’ work the machines had been doing.

Regulation is no further along. Under the EU’s labelling regime, an AI-written article can go unlabelled while a proofread email gets a marker, and American universities have nothing equivalent to point to at all.

Neither Stankova nor the San Francisco Standard has said whether a disclosure will be added to the piece. The argument it makes about UC admissions has not been withdrawn and remains, for the moment, the more consequential of the two disputes.

Get the TNW newsletter

Get the most important tech news in your inbox each week.