Professor Catches Most of His Class Using AI on a Midterm With One Hidden Prompt
The growing role of artificial intelligence in education is creating new opportunities for learning, but it is also raising serious concerns about academic honesty. A recent incident at Alcorn State University in Mississippi has reignited debate over how students use AI tools and how schools should respond when those tools cross the line from assistance to cheating.
History professor Jason Gibson drew widespread attention after sharing that the majority of his class appeared to have used AI chatbots to complete part of a midterm exam. According to Gibson, 32 out of 35 students submitted answers that suggested they had copied the exam prompt into an AI tool and relied on the generated response without carefully reviewing it.
The professor discovered the issue through a clever hidden instruction placed inside the exam directions. The concealed text told the AI chatbot to include the word “Madagascar” somewhere in the response in a way that made no sense.
Students who copied the full instructions into an AI chatbot unknowingly included that hidden command. As a result, their submitted answers randomly contained the word “Madagascar” in awkward or illogical places. The assignment asked students to compare technological advancements during the Industrial Revolution with modern technological developments, making the sudden appearance of “Madagascar” an obvious red flag.
What made the situation especially notable was not just the number of students caught, but how little effort some appeared to put into checking their work before submitting it. The strange wording suggested that many had trusted the AI-generated responses without reading them closely or thinking critically about whether the answer made sense.
In a follow-up explanation, Gibson said he allowed the students involved to challenge their grades. Only two students attempted to appeal. One student reportedly succeeded after explaining that she had seen the hidden instruction while using dark mode and believed it was an intentional part of the exam directions.
Gibson emphasized that his goal was not to humiliate students. Instead, he said the purpose was to encourage real learning, critical thinking, and responsible use of technology. The incident highlights a growing challenge for teachers and universities: AI tools can help students study, brainstorm, and better understand difficult subjects, but they can also make it easier to submit work that is not truly their own.
The case also reflects a larger shift happening across classrooms worldwide. As AI writing tools become more common, educators are being forced to rethink how they design assignments, measure understanding, and define academic integrity. Traditional take-home essays and online exams may now require new safeguards, clearer policies, and more emphasis on original thought.
AI in education is not necessarily a problem on its own. When used properly, it can support learning, explain complex topics, and help students improve their writing. But when students use AI to replace their own thinking, it undermines the purpose of education.
The Alcorn State University incident serves as a reminder that technology is changing the classroom faster than many institutions can adapt. Students need clear guidance on what counts as acceptable AI use, while educators need practical tools and policies to maintain fairness.
For now, Gibson’s hidden “Madagascar” prompt has become a memorable example of how easily careless AI use can be detected. More importantly, it has sparked a much-needed conversation about honesty, responsibility, and the future of learning in the age of artificial intelligence.
