Ethan Mollick reports tests where hidden prompts in letters and papers can manipulate LLM graders on older and smaller models but not most frontier models. The implication is evaluation pipelines need model-aware defenses.
“Can you prompt inject your way to an “A”?”
“people are inserting AI prompts into letters, CVs & papers.”
“It does on older & smaller models, but not on most frontier AI”
This finding is one of many signals tracked across Artificial Intelligence. The live feed updates every few hours with new authority voices, debates, and emerging ideas.
← Back to Artificial Intelligence