August 17, 2026, (Inside AI) — AI agents are now completing entire online courses for students, logging into learning platforms, watching lectures, taking quizzes, and writing papers. This shift goes far beyond pasting chatbot essays into a submission box. It outsources a full semester of academic work to software that can imitate a student's digital presence.
Instructors nationwide describe a wave of suspicious submissions that appeared suddenly, often polished beyond the student's demonstrated ability. Karen Costa, who has taught online for nearly 20 years, said she sensed an "influx" of AI-generated work last year. She recalled thinking, "Am I just interacting with bots here?" Costa, Karen Costa, online instructor and Northeastern University graduate.
The problem is most acute in asynchronous online courses, where students never meet professors or peers, even over video. More than half of American college students took at least one online class last year, up from about a third in 2019. Those classes are self-paced, with prerecorded lectures and digital submissions, making them uniquely vulnerable to agentic cheating.
Faculty say administrators are reluctant to punish suspected cheaters. AI detection software can produce false positives, leaving no concrete proof. Some administrators also fear losing paying students amid budget pressures. The result is an uneven response: some professors ignore the issue, others give poor marks, and a few allow unrestricted AI use.
Carol Sewell, who teaches in the California State University system, said she receives no guidance on discouraging AI use. "I'm not getting any guidance on how to dissuade their use," Sewell, Carol Sewell, gerontology instructor, California State University. "I'm getting opportunities to learn more about using it."
In tests by The New York Times, none of the three leading student AI tools, ChatGPT, Gemini, and Grammarly, refused a prompt asking for a paper written on a student's behalf. School-licensed versions can add barriers, but students bypass them with personal email addresses or second devices.
AI companies have not prevented agents from logging into learning platforms such as Canvas, Brightspace, and Blackboard. Those platforms can look for signals like time on task, but none offer a fail-safe block. Jason Weaver, Blackboard's chief of staff, said, "Our customers are making it clear that there has never been more pressure on academic integrity." Weaver, Jason Weaver, chief of staff, Blackboard.
Some educators want AI agents working inside online courses to identify themselves. AI firms have resisted. Perplexity, which makes an AI-powered browser popular with students, said requiring agents to identify themselves "would put the student's privacy and security at risk." Neither Google nor OpenAI responded to questions about their agents posing as students.
Cheating evolves faster than detection tools #
Experts warn that even newer forms of AI cheating are emerging. Wearable devices like glasses can let students request help while working in a secure browser. AI avatars can impersonate students over video. These tools make traditional proctoring and lockdown browsers less effective.
Some institutions are redesigning assignments to resist automation. In a recent online statistics class at the City University of New York's School of Professional Studies, participation grades depended partly on a discussion forum where students named the "clearest" and "muddiest" concepts from each unit and explained why. Ruru Rusmin, an assistant dean, said encouraging students to surface struggles removes the incentive to cheat.
Jordan Canzonetta, an associate professor at Lewis University in Illinois, said her biggest struggle is getting students to trust their own writing ability. "My biggest struggle is getting students to trust that they do have writing ability, that they have important things to say," Canzonetta, Jordan Canzonetta, associate professor, Lewis University.
Some faculty have called for government regulation of AI tools used by students. So far, tech companies have done little to discourage AI cheating. The issue affects both traditional and online classes, but it is harder to police when students complete work from home, far from professors' watchful eyes.
In the absence of technical solutions, some educators are adding in-person elements to online courses. Kathryn Kysar, a community college instructor, now requires in-person midterm and final exams for some classes, with a paper dictionary available at the front of the room. Marina Aminy, executive director of California Virtual Campus, said professors could offer flexible, in-person exam days without sacrificing access.