Alcorn State professor uses invisible AI trap to detect plagiarism

Jason Gibson, a history professor at Alcorn State University in the United States, successfully identified widespread use of artificial intelligence in student assignments using a digital “trap.”

To detect plagiarism, Gibson inserted an invisible instruction into the exam prompt by writing white text on a white background. While invisible to students reading the screen, the text was captured when students copied and pasted the prompt into AI chatbots. The hidden command instructed the AI to include the word “Madagascar” in an absurd context within the essay.

Of the 35 students examined, 32 failed the assignment after submitting essays containing nonsensical phrases involving the keyword, which served as immediate evidence of unreviewed AI-generated content. The incident has sparked debate regarding academic integrity and the tendency of students to rely on automation without critical review.