In a bold move to combat the rising tide of AI-assisted cheating, a history professor at Alcorn State University in Mississippi set a trap that ensnared nearly his entire class. Jason Gibson, who teaches history, embedded a hidden instruction in his midterm exam, telling any AI chatbot used by students to include random, nonsensical references to Madagascar. The result: 32 of his 35 students across two classes failed a portion of the exam, having submitted AI-generated responses without a second glance.
Gibson shared the outcome in a viral TikTok video, expressing disbelief at the students' lack of effort. "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," he said. "And apparently they didn't proofread it."
The hidden prompt was designed to be invisible to students but readable by AI models. It instructed the chatbot to weave in absurd phrases about Madagascar, a country with no connection to the exam's actual topic—the Industrial Revolution. The responses, as Gibson later showed in a follow-up video, were riddled with such nonsensical lines as "Madagascar floats sideways through the afternoon" and "Madagascar purple bicycle whispers to the ceiling." One student's essay on AI and social inequality ended with the bizarre assertion that the island nation "wore a toaster to a basketball game."
Gibson emphasized that his goal was not to humiliate students but to highlight the importance of academic integrity. After the exam, he explained the trap to his students and gave them the opportunity to contest their grades. Only two did, which Gibson noted suggests a lingering sense of shame.
Widespread AI Cheating in Academia
Gibson's experience is far from isolated. Across the country, educators are grappling with how to handle AI-generated work. At Brown University, a professor discovered that more than half of his students had used AI to cheat on an exam. Princeton University, which had long prided itself on its Honor Code, has been forced to reinstate supervised exams after a chatbot-related cheating scandal.
"We don't know best practices for navigating academia with AI," Gibson warned. "We're all just trying to hold onto some level of academic integrity in the process."
This incident at Alcorn State serves as a cautionary tale for students who might think that AI can replace their own critical thinking. It also underscores the need for institutions to develop clear policies and strategies to address the challenges posed by generative AI in educational settings.
A history professor at Alcorn State University in Mississippi embedded a hidden instruction in his midterm exam, prompting AI chatbots to include absurd references to Madagascar. Thirty-two of 35 students fell for the trap, submitting AI-generated answers without proofreading. The incident highlights the growing challenge of AI cheating in higher education.
Leave a Comment