AI Multi-agent safety gets Google’s $10 million test Google's $10 million multi-agent safety program starts small against the risk of interacting AI agents amplifying each other's failures. Data Today AI Desk · Jun 11, 2026
