A Berkeley Professor’s Op-Ed and AI Detection
Zvezdelina Stankova, a teaching professor of mathematics at UC Berkeley, penned an op-ed titled "I teach calculus at Berkeley. Some of my students can’t do middle school math" in the San Francisco Standard on August 15th. The article sparked a heated debate regarding UC Berkeley’s test-blind admissions policy.
Within days, the discussion shifted from the content of the op-ed to how it was written. A detector called Pangram, used by Substack to flag machine-written posts, flagged the article as roughly 33% AI-generated or AI-assisted. This raised questions about the role of artificial intelligence in its creation.
Stankova admitted to using AI for editing assistance but emphasized that the majority of the work was done by her, totaling around 80 hours. This defense, however, sparked further debate, especially considering the university’s policies on AI editing.
Some at Berkeley, like Hannes Bajohr, criticized the op-ed as deceptive, arguing for a clear disclosure when AI is used in writing. The original article, however, presents significant claims about the university’s calculus course enrollment and admissions disparities, which require independent verification.
The episode highlights a complex interplay between AI, academia, and intellectual integrity, especially as AI tools become more prevalent in writing and education. The detection methods themselves are not without controversy, adding another layer of complexity to the debate.