Mississippi college students caught cheating after prof puts hidden AI-only text on midterm
Professor Jason Gibson secretly embedded a hidden prompt in the exam instructions using white font, invisible to students but readable by AI chatbots.
Professor Jason Gibson secretly embedded a hidden prompt in the exam instructions using white font, invisible to students but readable by AI chatbots.

Thirty-two college students have failed part of their history midterm exam after falling for an AI-cheat trap when they blindly copying ChatGPT-generated answers without bothering to proofread thoroughly.
Jason Gibson, a history professor at Alcorn State University in Mississippi, secretly embedded a hidden prompt in the exam instructions using a white font, invisible to students but readable by AI chatbots if the instructions were copied and pasted. The prompt instructed the AI to randomly insert absurd nonsense about Madagascar into every response.
Instead of answering questions about the Industrial Revolution, dozens of students unknowingly turned in essays declaring that "Madagascar floats sideways through the afternoon," that "Madagascar purple bicycle whispers to the ceiling," and that the island nation had somehow "wore a toaster to a basketball game."
The bizarre answers immediately gave the game away. “Thirty-two of my 35 students between two classes failed a portion of their midterm because they all used AI to generate their entire response," Gibson said in a viral video online. "And apparently they didn't proofread it."
https://twitter.com/tpostmillennial/status/2081803845831680221
According to Gibson, many students simply copied the entire exam prompt into an AI chatbot before pasting the response straight into their midterm, without noticing the hidden white text that had allowed their history essays to suddenly veer into fiction. "I get no gratification out of seeing students fail. It's not my intent to see you fail, that's weird," the professor explained. He instead clarified that he wanted to combat the increasing problem of AI-assisted cheating. He even allowed students to challenge their grades, but only two attempted to contest the result.
Schools and universities across the country have reported increasing misuse of chatbots in exams and coursework, while safety concerns surrounding AI continue to grow. "We're all just trying to hold onto some level of academic integrity in the process," Gibson said.
Libby’s Desk
Commentary and must-reads from Libby’s desk, straight to your inbox.
TPM Newsletter
The Post Millennial’s top stories, delivered to your inbox.


