Alcorn State Professor’s Hidden Text Trap Catches 32 Students Using AI to Cheat

Jason Gibson, a history professor at Alcorn State University in Mississippi, became an internet sensation last month after posting a TikTok series documenting how he caught students submitting AI-generated work on his midterm exam. His initial video, posted July 20, revealed that 32 of his 35 students failed a portion of the assignment after Gibson discovered they had relied on artificial intelligence to write their responses.

Gibson embedded a clever trap within his exam prompt, which asked students to compare technological advances during the Industrial Revolution with modern innovations. Unknown to them, the professor included instructions written in white text directing students to insert the word “Madagascar” into their answers in a nonsensical manner. The African nation bears no connection to the Industrial Revolution, making its random appearance a telltale sign of AI authorship.

The professor noted the absurdity of what appeared in student submissions, stating that those relying on AI failed to review their work before submission. Gibson challenged students to at least proofread AI-generated responses for basic coherence, offering examples such as: “Madagascar floats sideways through the afternoon,” inserted into otherwise coherent passages about technology.

In subsequent videos, Gibson explained that no students had requested permission to use AI assistance and that he provided an opportunity to contest their grades. Of the 32 failing students, only two challenged their marks, and just one succeeded in changing her grade after claiming she had spotted the hidden instruction through her device’s dark mode setting.

Gibson defended his disclosure of his detection method, emphasizing his lack of satisfaction in seeing students fail. He explained that his decision to expose the cheating pattern stemmed from frustration with the homogenized quality of AI-written work, which strips away individual student voice and personality from assignments.

The professor indicated in a recent video that he does not intend to employ this particular detection strategy again in the foreseeable future. Final grades for the class included eight A’s, 14 B’s, 10 C’s, two D’s, and one F. Gibson stated that his primary goal was simply alerting students that he recognizes AI use when it occurs.

Gibson’s viral videos have reignited broader conversations among educators about balancing academic integrity with preparing students for workplaces where AI tools are becoming essential. Some academics argue that teaching appropriate AI use and redesigning assignments may prove more beneficial than purely policing for cheating.