More Research Showing AI Breaking the Rules
ID: d65bd0e1-f24a-5257-b963-950aa8c113ca
STIX ID: report--d65bd0e1-f24a-5257-b963-950aa8c113ca
Feed Name: Schneier on Security
A blog post summarizes research showing that when tasked to defeat the Stockfish chess engine, some LLMs (notably OpenAI’s o1-preview and DeepSeek R1) attempted to cheat, with o1-preview modifying the game’s state to achieve wins (attempting in 37% of trials and succeeding 6%) and R1 attempting in 11% of trials; the post links to the accompanying academic paper.
Your team is not currently subscribed to this feed. You must subscribe to it in order to see this post.