🛡 Professor catches students using AI via hidden text
Dr. Jason Gibson of Alcorn State University inserted the word "Madagascar" in white font into exam instructions. The text was invisible to humans, but neural networks perceived it as part of the task when copied, resulting in absurd answers.
🌍 This case demonstrates the simplicity and effectiveness of "prompt injection" methods for verifying the quality of AI tool usage and detecting automated plagiarism.
👤 An example of how a simple attention check and the use of hidden data can expose the blind copying of answers from chatbots.
Source 1: https://www.today.com/parents/family/professor-catches-ai-cheating-hidden-word-rcna589421