Berkeley professor faces controversy over AI use in critique of student preparedness

A mathematics professor at the University of California, Berkeley, has sparked controversy after admitting to using AI tools to edit an article criticizing students' weak foundational math skills. Zvezdelina Stankova, a professor at the university, published a 2,000-word commentary in the San Francisco Chronicle, arguing that some of her students lacked basic math knowledge typically taught in middle school. She attributed the issue to California universities' decision to stop requiring standardized test scores like the SAT in admissions. Weeks later, a student at Berkeley, Francis Luo, used the AI detection tool Pangram to analyze the article and found that 33% of its content was likely generated or assisted by AI. Stankova confirmed she had used AI to help with editing but emphasized that the article was the result of hundreds of hours of manual work, with approximately 80 hours of direct contribution from her. Stankova explained that she used AI to gather and analyze data related to the initiative, but all the analysis was conducted by her team members. The San Francisco Chronicle responded by reiterating its policy that every article must have a human author responsible for every word, stating that the article reflected Stankova’s and her colleagues’ original research and work. The incident has fueled broader debates about the use of AI in academic writing. Critics argued that it was ironic for an article discussing student preparedness to rely on AI assistance, while supporters defended the use of AI as not a major issue and argued that the debate had shifted focus from the article’s main topic. Stankova stated that the use of AI was unrelated to the team’s efforts to push for the reinstatement of standardized testing in admissions. The controversy comes amid ongoing discussions about whether California universities should reinstate SAT and ACT requirements. In 2020, the University of California system voted to gradually eliminate the requirement for applicants to submit standardized test scores, a decision mirrored by other top universities like Harvard, Stanford, MIT, and Yale. These institutions have since reinstated the requirement. Stankova’s article cited data from her and her colleagues showing that the number of students entering calculus courses with severe foundational gaps had tripled since the removal of standardized testing. She also noted that she had co-signed a June open letter with over 3,000 UC faculty members supporting the reinstatement of standardized testing, and that the letter itself had also used AI tools for editing. The UC system has an AI committee that assists with AI projects and resources, but committee member Camille Krittenen stated that the university had not established specific policies regarding the use of AI by faculty in personal research or commentaries. She also pointed out that most AI detection tools remain unreliable and lack detailed judgment capabilities. Pangram, the tool used to detect AI-generated content, claims to have an accuracy rate of 99.66%, with an average of one false positive for every 24,000 documents scanned.