The clue documented prompt copying
The nonsensical word was embedded in white text inside the assignment. Students who copied the entire prompt into a model received outputs that obeyed the hidden instruction, making the misuse visible in the final answer. Two students reportedly challenged the grade, and one succeeded after saying dark mode exposed the text and made it appear to be a legitimate requirement.
That appeals process is not a footnote. Any automated or trap-based integrity system can produce ambiguous evidence, and students need a meaningful opportunity to show how they worked.
Detection cannot carry the curriculum
The professor described becoming weary of generic, personality-free responses and said the purpose was to signal that the pattern was visible—not to take pleasure in failure. Other educators quoted in the report argued that graduates will need AI at work, making blanket avoidance unrealistic.
Assignments can respond by requiring source checks, drafts, reflection on model errors, live explanation, or application to unfamiliar material. Those designs do more than deter cheating: they turn responsible AI use into something a student must understand rather than hide.
Go to the source
Read the evidence behind this analysis. External links open in a new tab.
Fox News — Mississippi professor reveals hidden way he used to expose AI cheating
