A recent survey found that only 42 percent of Americans express confidence in higher education. Meanwhile, a USA Today opinion piece by Natalie Sandoval, a 2024 Columbia University graduate, examining the rise of AI-assisted cheating on college campuses.
Proft welcomed Sandoval, an opinion fellow with USA Today through the Intercollegiate Studies Institute, to discuss her reporting. Sandoval described what she called the largest known AI-assisted cheating scandal in the Ivy League, involving a Brown University economics course in which a professor administered a take-home midterm for the first time. Where historical exam scores in the course had typically ranged between 65 and 80 percent, the take-home exam produced an average score of 96 percent, prompting suspicion of widespread cheating. When the professor switched the final exam to an in-person format, the average score dropped to roughly 49 percent, a result Sandoval said illustrated both the prevalence of AI-assisted cheating and a deeper erosion of genuine subject mastery among students relying on the technology.
Sandoval said cheating culture on elite campuses extends well beyond AI tools, citing separate reporting indicating that nearly 40 percent of students at Stanford carry some form of disability accommodation, which she said can confer benefits including extended testing time, preferential housing placement, and expanded dining options. She argued that such accommodations are sometimes sought opportunistically rather than reflecting genuine need, and said the broader culture of credential embellishment often begins well before college, particularly among students navigating competitive preparatory school and admissions environments.
Proft connected the discussion to broader debates over college admissions practices following the Supreme Court’s ruling against race-conscious admissions in the Harvard case, citing data showing that a 1550 SAT score afforded white or Asian applicants a 76 percent acceptance probability at Columbia in the 1990s compared to a 16 percent probability today. Sandoval agreed that many students enter the admissions process aware that evaluation criteria extend well beyond pure academic merit, noting that personal essays referencing identity or background often serve as an alternative avenue for demonstrating diversity credentials following restrictions on explicit demographic consideration.
Sandoval placed responsibility for the Brown cheating incident squarely on students rather than the professor, arguing that requiring in-person, handwritten exams remains one of the few reliable ways to limit AI-assisted cheating, while acknowledging that some academic disciplines lend themselves to this approach more easily than others. She said survey data suggests roughly three-quarters of Brown faculty report concern about AI-related academic dishonesty, and that many professors at her own alma mater struggled throughout her final semesters to establish consistent policies governing permissible AI use.
Asked about the broader dynamic between students and faculty on elite campuses, including concerns that professors may avoid confronting disruptive behavior for fear of backlash, Sandoval said she did not personally witness such dynamics extending into AI-related enforcement, though she acknowledged broader reporting describing professors facing public pressure campaigns from students. She said that during her time at Columbia, classroom interactions were generally respectful despite the presence of strongly progressive viewpoints among both students and faculty, and characterized the university’s most vocal, confrontational elements as representing a vocal minority rather than the broader student body, while noting that many students hold firmly established ideological assumptions they are resistant to reconsidering when challenged.