A cheating crisis is unfolding in American universities due to the growing integration of AI in education. Extreme classroom practices aimed at preventing deception are increasingly common, leading to false accusations against students. The definition of cheating is changing, according to professors, students, and academic integrity experts.
Increasing Measures to Prevent Cheating
At UCLA, a sociology class required students to use mirrors to reflect their workspaces and activate their laptop cameras for online test monitoring. Another course required students to take video exams with their arms crossed to prevent typing into AI platforms. A Los Angeles attorney handling students’ disciplinary cases noted that AI accusations constitute about 35% of her education caseload and are rising quickly. Some professors have reported over half a class for AI violations.
Ambiguity in Cheating Definitions
Differentiating cheating has become complex, with AI policies varying widely. Some faculty allow AI source citations in drafts, while others, like UC Berkeley law school, have banned nearly all AI use. Lee Rainie from Elon University highlights trust as the root issue, with students suspecting faculty of undisclosed AI use and vice versa.
Prevalence of AI Use in Classrooms
Igor Chirikov from UC Berkeley surveyed over 95,000 students at 20 public universities and found that two-thirds had used AI for classwork, with one-third using it regularly. About 9% admitted to using AI to cheat, by writing papers or completing assignments contrary to rules. Chirikov noted a substantial increase in AI use over the past two years, now at 80%.
The Challenge of Proving Cheating
University conduct officers express difficulty in proving AI-based cheating, and academic AI-defense lawyers are emerging to support accused students. Some students are falsely accused due to professors’ suspicions or flawed detection software. Turnitin, a common AI-checker, claims a less than 1% error rate but acknowledges the risk of false positives.
Extreme Monitoring Practices
UCLA’s Titi Olotu experienced unusual AI monitoring in an online class, leading her to drop the course. She felt restricted by the monitoring, viewing any movement as suspect. While lock-down browsers and shared screen videos are common, mirrors and body movement restrictions are viewed as extreme.
Students Building Defenses Against Accusations
Students like Ivan Ornelas and Aldan Creo take measures to avoid AI accusations, focusing on creating detailed version histories of their work and altering writing style to evade detection. Attorney Adrienne Hahn advises students to gather evidence, such as study notes and version histories, to defend against false accusations.
Focus on System Preparedness
Tricia Bertram Gallant from UC San Diego emphasizes the inadequacy of colleges to handle AI use in classrooms, rather than the notion of widespread false accusations. While some instructors report AI violations, confusion persists over AI use versus academic misconduct.
Efforts to Detect Cheating
Adam Kaiserman at College of the Canyons initially reported a student based on AI scanner results but faced challenges when more advanced detection tools disagreed. Kaiserman altered his teaching methods to deter cheating, such as rewarding students for placing phones at the room’s front for extra credit. Advanced assignments have been reconsidered due to AI capabilities.
Gallant argues that the real challenge is the struggle to adapt teaching methods to new technology, rather than false accusations. She highlights the need for supervised and observed academic work to maintain integrity.
Leave a Reply