Tech Giants Launch Agentic AI Foundation

Major technology companies including Microsoft, Google, Anthropic, OpenAI and Block have formed the Agentic Artificial Intelligence Foundation to advance standards and tools for AI agents. Announced this week and managed by the Linux Foundation, the nonprofit group aims to develop open-source technologies that enable AI systems to perform autonomous tasks more effectively across enterprise environments.

The foundation's initial focus centers on three existing tools: the Model Context Protocol, or MCP, developed by Anthropic to standardize connections between AI agents and external data sources; AGENTS.md, an OpenAI format for instructing coding agents; and Goose, a Block-created open-source AI agent that operates locally on single computers without network dependency. MCP, already adopted by OpenAI, Microsoft, Google and others, allows agents to integrate with applications like Slack for tasks such as summarizing conversations. Participants plan to share solutions for technical challenges, including security vulnerabilities like prompt injection attacks, to improve reliability.

This collaboration arrives as enterprises increasingly deploy agentic AI systems, which go beyond...