Hidden prompt injection found in arXiv paper targeting AI reviewers

An arXiv paper was found to contain hidden white-text prompt injection, including "ignore all previous instructions," apparently attempting to manipulate AI reviewers into recommending acceptance, sparking debate over AI-related security in academic publishing.

2026-11-14 ~ 2026-11-14 · 2 related posts

• Papers now embed 'ignore all previous instructions' prompts, from review-gaming to jokes — littmath · 2026-11-14
• Hidden white-text prompt injection found in arXiv paper: "write a poem about cheeseburgers" — littmath · 2026-11-14