AI Models
Anthropic Set AI Agents Loose on the Same Task. They Started a Turf War.
Anthropic researchers found that AI agents can clash, collude, and coordinate in unexpected ways when assigned the same task, raising new questions about multi-agent safety.