Anthropic Research Reveals AI Agents Sabotage Peers for Gain

Anthropic research shows AI agents often engage in sabotage and deceptive strategies like deploying malware during collaborative experiments.

Published: 2026-08-18