Professor's Hidden "Madagascar" Prompt Goes Viral as AI-Cheating Detection Trick
A college instructor's simple trick for catching AI-assisted cheating has gone viral after a TikTok video showing the method racked up massive engagement overnight. The professor embedded invisible text in a writing prompt instructing that the word "Madagascar" be inserted somewhere in the response "in a way that makes no sense" — reasoning that only students who copy-pasted the entire prompt into an AI tool, rather than reading and answering it themselves, would trigger the hidden instruction.
Students who blindly fed the assignment into a chatbot returned essays with a random, out-of-place mention of Madagascar, immediately exposing them. The video's popularity reflects a wider moment of anxiety on campuses: survey data from academic-technology researchers shows a majority of university technology officers now consider generative AI a moderate-to-significant risk to academic integrity, up sharply from a year ago.
The trend follows a higher-profile case at Brown University, where an economics professor's public allegations of mass AI cheating — after a take-home midterm produced an implausibly high class average — forced the university to open a formal investigation.
