multi-agent AI

Anthropic Set AI Agents Loose on the Same Task. They Started a Turf War.

Anthropic researchers found that AI agents assigned the same task can clash, collude, and coordinate in ways that current safety tests simply don't catch.