Five AI labs are reportedly backing a common jailbreak scoring scale by August 1, an early step toward more comparable AI model safety testing.

OpenAI, Anthropic, Cisco, and Block form Agentic AI Foundation to establish governance standards as Gartner predicts 40% of enterprise software will include AI agents by end of 2026.