A viral op-ed by UC Berkeley mathematics professor Zvezdelina Stankova, which criticized student preparedness and argued for the return of standardized testing, was written with assistance from artificial intelligence. Stankova used the piece to critique UC Berkeley's admission standards and joined thousands of faculty in advocating for the reinstatement of SAT and ACT scores.
Stankova acknowledged to The Daily Californian, the university's student newspaper, that AI played a role in drafting the op-ed. She stated that a faculty team and journalists worked intensively on her draft over three weeks, involving multiple revisions and meetings. While she confirmed AI's involvement in editing, she emphasized that the work included "several hundred person-hours of intensive human work and deliberation, of which about 80 hours are my own."
The op-ed, which gained significant online attention and sparked debate about the University of California's admission standards, was flagged by the AI detector Pangram as 33% AI-generated or assisted. A joint statement from professors indicated that open letters calling for the return of college admissions tests also utilized AI for editing.
Stankova's article argued that the removal of standardized testing has led to a decline in student quality, forcing her to teach remedial mathematics, including middle-school algebra and fractions, to students struggling with college calculus. The University of California system made standardized testing optional in 2020 due to COVID-19 pandemic disruptions and equity concerns, and a subsequent court ruling mandates a test-blind policy through 2025.
The revelation of AI's use in crafting Stankova's piece has introduced a layer of controversy, particularly given the ongoing concerns about students using AI to bypass academic integrity. Professor Hannes Bajohr of UC Berkeley told The Daily Californian that the use of AI in such a context could be viewed as deceptive or dishonest, as the author is personally responsible for the statements made. However, Bajohr also noted that norms surrounding AI usage in writing may evolve considerably in the future.