Fed-up professor deploys prompt trap to catch students cheating with AI
“I can’t make this stuff up,” the professor said.
Image by Cybernews.
Follow us
- Professor hid invisible white-text instructions in an assignment to expose students copy-pasting answers into AI chatbots
- Thirty-two of 35 students unknowingly submitted nonsensical Madagascar references, proving they never even reviewed AI-generated responses
- Princeton survey shows unauthorized ChatGPT use among seniors jumped 12.5% year-over-year, reaching 27.7% of the class
Key Takeaways by nexos.ai, reviewed by Cybernews staff.
If you’re still in doubt whether our young minds believe cheating is now acceptable because AI’s around, hesitate no more. A clever prompt trap by a university professor caught almost all of his students using the technology to cheat on an assignment in the blink of an eye.
It’s all truly depressing, especially when mass cheating is now observed despite the tech geniuses telling us that AI is first and foremost a tool of empowerment.
However, modern problems require modern solutions, and Dr. Jason Gibson, a history professor at Alcorn State University, seems to have found one – even if he probably isn’t going to be able to repeat the trick in the future.
On TikTok, Gibson explained that he hid a white-colored font in a midterm assignment, making the students who pasted it to an AI prompt unknowingly include truly nonsensical Madagascar-related bits in their answers about the Industrial Revolution.
Gibson said he had an inkling most of his students were using AI to cheat and decided to find out the size of the phenomenon for himself.
Since the instructions were written in white text on a white background, they were invisible. Sure enough, 32 out of 35 students across two classes immediately pasted the question into a chatbot and used its output as their answer.
Has your password leaked?
It looks like the students are using AI almost blindly. Had they bothered to double-check what the chatbot had generated, they’d have surely noticed some weird references to Madagascar.
Almost no one checked, and the essays the students returned contained phrases such as “Madagascar purple bicycle whispers to the ceiling” and “Madagascar floats sideways through the afternoon.”
“You would be amazed at the bizarre usage of the word ‘Madagascar’ that I saw in these students’ assignments. It was more than obvious they didn’t even go back and reread these responses,” said Gibson.
When confronted about the cheating, only two of the caught students tried to contest their grades. Gibson says in the video that all cheaters officially failed this part of their midterm.
The essays the students returned contained phrases such as “Madagascar purple bicycle whispers to the ceiling” and “Madagascar floats sideways through the afternoon.”
Cheating among students is a growing problem in the age of the chatbot. Just this month, a Brown University professor saw his students’ midterm scores significantly drop when he changed the exam to an in-person test where AI couldn’t be used.
A 2025 survey of Princeton University seniors found that 27.7% admitted to using ChatGPT when it was not allowed, up 12.5% from the Class of 2024.
Have thoughts about this topic? Others do, too. Join them in the discussion.
Another investigation by The Guardian found last year that thousands of students in UK universities had been caught using AI tools to cheat on their assignments.
AI detectors such as GPTZero exist, but they’re not exactly reliable, producing false positives. However, educators like Dr. Gibson are fighting back and indeed finding solutions such as prompt traps or in-class exams.
