Fraudsters Are Enrolling Students Who Do Not Exist, Then Letting AI Do the Homework
US community colleges are dealing with a new kind of scam: fake enrolments to collect financial aid, with chatbots quietly completing the coursework that keeps the money flowing.
There is a scam running through US community colleges that only works because AI got good. Fraudsters enrol students who do not exist, collect the financial aid attached to those enrolments, and then use chatbots to complete just enough coursework to keep the enrolment looking legitimate. Reporting in The New Yorker lays out how it works and how long it took anyone to notice.
Professor David Song at East Los Angeles College says he spotted it a few years ago, and the tell was demographic rather than technical. Students with generic Anglo-Saxon names started appearing in his history course at a college whose student body is largely Latino and Asian. When he checked, their claims about having taken courses on Asian American topics did not hold up. The abuse concentrates in asynchronous online courses, the kind where nobody ever appears on camera and a name on a roster is the only evidence a person exists.
The same reporting captures a second, more ordinary problem sitting on top of the fraud. Song allows AI in his courses, but requires students to label it and be honest about what they used. Nobody respects those rules, he says, “even with the content that’s blatantly generated.” History professor David Roach estimates that more than half his students use AI for papers. His question is the interesting one, and he asks it without much bitterness: “Was it always the case that half of our students would cheat if it were easy enough?”
That question is worth sitting with, because it points at what actually changed. Enrolment fraud and coursework cheating both existed before 2022. What AI removed was the cost. Faking a semester of assignments used to require either real effort or paying someone, which made the scam small enough to ignore. Now the marginal cost of a plausible essay is near zero, and any system whose integrity quietly depended on effort being expensive has a problem. The colleges are not being outsmarted so much as having an old, unstated assumption pulled out from under them.
What this means for you: For most readers, directly, nothing. But this is the clearest everyday illustration of a pattern worth carrying around: AI does not usually invent new crimes, it makes existing ones cheap enough to run at scale. If you work anywhere that verifies people or reviews submitted work, whether that is a school, a grants office, a hiring pipeline or a small association handing out memberships, the check that used to work because faking it was tedious probably does not work any more. The fix is rarely a better AI detector, which tends to be unreliable and unfair to honest people. It is usually a moment of real human contact somewhere in the process: a conversation, a live task, a face.
Sources
Source: https://www.newyorker.com/news/fault-lines/the-despair-of-the-professor-in-the-age-of-ai
Google DeepMind Is Becoming Just Another Google Department, and Its Founder May Be on the Way Out
Demis Hassabis has been moved to chairman and may leave for good. Gemini development is moving to the Bay Area. Two very different readings of what that means for Google.
