U.S. Students Use AI Agents to Attend Classes, Sparking Education Concerns

Recently, growing attention has been drawn to reports that some students in the United States are using AI agents to complete online courses. As AI agent models become more widespread, certain students are no longer merely using chatbots to look up answers; instead, they are delegating entire courses to AI agents operating in the background—tasks include logging into learning platforms like Canvas and Blackboard to take quizzes, watching lecture videos, writing essays, and participating in discussion forums, all while the students themselves engage in activities unrelated to their coursework. In-person classes can verify students’ mastery of material through exams, oral presentations, and other direct assessments. However, online courses—conducted primarily on-screen—pose significant challenges for implementing equivalent oversight. Currently, widely used AI tools such as ChatGPT, Gemini, and Grammarly lack mechanisms to prevent such misuse and have not responded to educators’ calls for AI systems to self-identify when operating on educational platforms. In response, some universities have begun taking action. The University of Chicago Law School recently introduced a new “AI strategy” prohibiting first-year students from using phones, tablets, and laptops during class. Meanwhile, Princeton University, following an AI-related cheating incident, abandoned its century-old “honor code” tradition, discontinuing unsupervised exams that previously relied on student self-regulation.