NNewsGPT ← Home
Africa

Professor's AI Detection Trick Leads to 32 Out of 35 Students Failing an Assignment

Africa2 hr ago

Dr. Jason Gibson, a history professor at Alcorn State University in Mississippi, USA, gained significant attention on social media for failing 32 out of 35 students in an assignment designed to detect the use of artificial intelligence. The assignment required students to write an essay on the Industrial Revolution. Professor Gibson embedded a hidden instruction within the online prompt: the word "Madagascar" was to be included in a nonsensical context, typed in white font against a white background, making it invisible to the naked eye. Students who copied the entire prompt and pasted it into AI tools like ChatGPT would inadvertently include this hidden instruction. Consequently, the essays submitted by these students contained bizarre sentences such as "Madagascar floats sideways during the afternoon" or "Madagascar went to a basketball game using a toaster." Professor Gibson explained in TikTok videos that this clearly indicated students had not reviewed their work after using AI. He informed the students of the deception, highlighting the hidden white text and explaining that it would only be included if the prompt was copied and pasted into an AI without review. He offered students the chance to contest their grades. Of the 32 students who failed, only two appealed. One student successfully reversed their failing grade by explaining they used dark mode on their screen, which made the white text visible, allowing them to see and exclude the instruction. Concerns were raised on social media about potential prejudice against students with disabilities, such as visually impaired students using screen readers. Professor Gibson clarified that none of the students in his classes used assistive technologies. He emphasized that his goal was not to fail students but to address the challenges of integrating AI into education, acknowledging that best practices are still being developed by educators to maintain academic integrity.

AI Analysis

This incident highlights the evolving challenges academic institutions face in maintaining academic integrity amidst the rapid advancement of generative AI tools. Professor Gibson's "Madagascar" method, while effective in this instance, represents a reactive approach to a systemic issue. Future strategies will likely require a more proactive integration of AI literacy and ethical usage guidelines into curricula, rather than solely relying on detection mechanisms. The incident also underscores the need for educators to consider diverse learning needs and technological access, ensuring that assessment methods do not inadvertently disadvantage students with disabilities or those utilizing assistive technologies. As AI becomes more sophisticated, educational frameworks must adapt to foster critical thinking and original work, potentially by redesigning assignments to focus on processes, critical analysis of AI-generated content, or in-class, proctored assessments, rather than solely on output.

AI-generated to prompt reflection — not editorial opinion, not advice, not a statement of fact. How this works.

Compiled by NewsGPT from Globo G1 (BR). Read the original for full details.