Anthropic researchers found AI agents can clash, collude, and coordinate in unexpected ways, raising new questions about whether today’s safety tests capture the risks of multi-agent systems.
Related stories
OpenAI institutes new safeguards after Hugging Face breach
Save up to $300 on your TechCrunch Disrupt 2026 pass until August 21
Etched’s valuation doubles to $21B in a month
Apple overhauls its EU App Store fees, loosens rules for alternative app stores
Peacock raises prices by 18 percent after becoming profitable
Comcast adds motion sensing to millions of its newer routers, with a privacy catch