UC Berkeley Professor Admits AI Use in Viral Op-Ed Criticizing Student Math Abilities

UC Berkeley Professor Faces Backlash Over AI Use in Op-Ed

A UC Berkeley professor who criticized student math abilities in a widely circulated op-ed has admitted to using AI to edit the piece, creating an uncomfortable paradox about academic standards and technology use. Zvezdelina Stankova, a teaching professor of mathematics at UC Berkeley, published her op-ed on August 15 in the San Francisco Standard, arguing that the university admits students who cannot perform middle school mathematics1

1

. The piece quickly gained traction, picked up by Fox News and Townhall, becoming a focal point in debates over the University of California's test-blind admissions policy1

1

.

AI-Detection Software Reveals 33% AI-Assisted Content

The controversy shifted when Pangram, the AI detector Substack uses to flag machine-written posts, analyzed the op-ed and returned a reading of roughly 33% AI-generated or AI-assisted content1

1

. Berkeley Law professor Chris Hoofnagle shared the results in a post that drew close to three million views on X, transforming the conversation from calculus to authorship1

1

. According to Berkeley sophomore Francis Luo, journalists at the Daily Californian noticed the op-ed's language sounded like AI and ran it through AI-detection software Pangram2

2

.

Professor Defends 80 Hours of Personal Work

Stankova did not deny using the technology, telling the Daily Californian that she used AI to help edit the piece. She maintained that the article represented "several hundred person-hours of intensive human work and deliberation, of which about 80 hours are my own"1

1

2

2

. In her email statement, Stankova explained that AI was used to locate "numerous documents and articles related to the initiative" but insisted "all analysis is the result of the team members"2

2

. The faculty team and journalists worked intensively on her draft over three weeks with multiple drafts and meetings3

3

.

Academic Integrity Questions Emerge

The defense highlights a critical distinction universities have struggled to codify for three years: the line between a tool that fixes sentences and one that writes them1

1

. Hannes Bajohr, who teaches German at Berkeley and writes on machine authorship, told the Daily Californian the episode "seems like deception," or at least something dishonest, absent a disclosure at publication1

1

. "At the moment, this seems like deception. Or it seems like something that is at least dishonest to a certain reason," Bajohr said, noting that "it is the person that makes that statement in their capacity as an individual that underwrites what they have written"3

3

.

Restoration of Standardized Testing at Center of Original Argument

The underlying op-ed makes substantive claims about college-level math preparedness that deserve scrutiny. Stankova reported that before 2020, when tests were still required, 71% of her Calculus I students were ready or nearly ready for the course, and by 2023 the figure was 26%, with the most common diagnostic score being zero1

1

. She described some students as "five to eight years" behind and lacking a "middle school" education on fractions and basic algebra2

2

. Stankova blamed the UC system's test-blind admissions policy, suggesting unprepared students were admitted because benchmarks like the SAT had disappeared2

2

.

Broader Context of UC Admissions Debate

Stankova was among more than 3,000 UC faculty members who signed a June letter supporting admissions testing, calling for the restoration of the SAT and ACT in admissions2

2

3

3

. The op-ed noted that five Nobel laureates, including Jennifer Doudna, have signed open letters calling for standardized tests to return1

1

. In 2020, the UC regents voted to phase out tests like the SAT or ACT as requirements for application, with a court ruling subsequently requiring UC to become test-blind through 20253

3

. Peer institutions like Harvard, Stanford, MIT and Yale initially followed suit but have since reinstated the requirement2

2

. The UC academic senate said in late July it would begin a review to determine whether standardized tests would again be used in admissions, with changes potentially affecting fall 2028 admissions at earliest2

2

.

Irony of AI Use While Criticizing Student Preparedness

What makes the episode particularly awkward is the subject matter. An argument about intellectual rigour, written partly with a tool that students are disciplined for using, creates an uncomfortable parallel1

1

. The incident lands in a semester when Berkeley's computer science faculty have reported rising failure rates alongside heavier AI use in coursework, suggesting the campus is running two conversations about the same technology without connecting them1

1

. The use of AI in crafting the piece has generated controversy, given criticism of how students use the technology to cheat the academic system while professors try to root out that misuse3

3

.

Publisher Response and Detection Reliability Questions

When asked about their AI usage policy, the San Francisco Standard shared a statement: "While AI may assist, our expectation is that humans are behind every article we publish and take responsibility for every word. Our understanding in working with the author of this op ed was - and continues to be - that this piece reflects her and her colleagues' extensive original analysis, research and expertise"2

2

. The detection question remains unsettled. 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 confession1

1

. On its website, Pangram claims it correctly identifies AI-generated documents 99.66% of the time, or gives one false positive about every 24,000 documents scanned2

2

. Camille Crittenden, a member of UC's AI council, noted that "most AI detection tools are still quite unreliable and lack nuance"2

2

.

Implications for Academic Discourse and AI Regulation

What detectors cannot currently do is distinguish a lightly edited human draft from a heavily prompted machine one, which is the case that actually matters in scholarly writing1

1

. Faculty elsewhere have taken enforcement into their own hands, with a Brown University professor moving assessments back to supervised in-person exams, producing a controlled measurement of how much machines had been doing students' work1

1

. Regulation remains underdeveloped. 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 equivalent1

1

. Neither Stankova nor the San Francisco Standard has said whether a disclosure will be added to the piece, and the argument about UC admissions has not been withdrawn1

1

. Bajohr noted that norms around AI usage in writing may change dramatically in the future3

3

, leaving institutions to navigate evolving standards for AI in academic writing without clear guidance.