Skip to main content

Professor's Hidden 'Madagascar' Trap Exposes AI Cheating in Exam

A professor's ingenious method to catch students using AI to cheat on an exam has taken the internet by storm, raising serious questions about academic integrity in the age of easily accessible artificial intelligence.

Professor Jason Gibson recently shared a TikTok video that garnered over ten million views, detailing how he exposed 32 out of 35 students in his two classes for using AI-generated answers on a midterm exam. The students didn't even bother to read the AI's output before submitting it, leading to a collective failure that was both shocking and, to many, amusing.

Gibson's anti-cheating strategy was clever and elegant. He embedded a hidden instruction in the exam, written in white text that was invisible to the naked eye. When students copied the questions into an AI tool, this hidden prompt caused the AI to generate a meaningless sentence containing the word "Madagascar." As Gibson graded the papers, the sudden appearance of "Madagascar" in 32 answer sheets immediately revealed the cheating.

After the exam, Gibson publicly explained his screening method and offered students a chance to appeal. Only two out of the 35 students chose to do so. In his video, Gibson shared some of the absurd "Madagascar" sentences generated by the AI, which sparked a heated discussion on social media about AI and academic honesty.

This incident highlights a growing concern for educators worldwide: as AI tools become more accessible, traditional exam formats are increasingly vulnerable to cheating. Students no longer even need to read the AI's output before submitting it. In Gibson's case, if any student had glanced at those strange paragraphs about Madagascar, the entire class might not have collectively failed.

The question now is not just how to stop students from using AI, but what happens when the basic willingness to check AI answers is disappearing. How much of the educational foundation remains when students blindly trust machine-generated content?

Key Points

  • Professor Jason Gibson used a hidden prompt in an exam to catch AI cheaters.
  • The prompt caused AI tools to generate sentences containing "Madagascar."
  • 32 out of 35 students fell for the trap and failed the exam.
  • Only two students appealed after the professor explained the screening method.
  • The incident has sparked a debate on academic integrity in the age of AI.