OpenAI, Anthropic, and Block Join Forces to Standardize AI Agents
OpenAI, Anthropic, and Block Join Forces to Standardize AI Agents
As artificial intelligence evolves from simple chatbots to systems capable of executing actions, major players in the tech industry are taking steps to ensure a future of compatibility and openness. OpenAI, Anthropic, and Block have partnered with the Linux Foundation to launch the Agentic AI Foundation (AAIF), a group dedicated to creating standards and open-source tools for the emerging AI agent landscape.
Building the Framework for Compatibility
The AAIF aims to prevent the fragmentation of AI agents into proprietary, incompatible systems by acting as a neutral hub for open-source projects. Initial contributions from Anthropic, Block, and OpenAI include technologies designed to form the foundation of interoperable AI infrastructure.
Anthropic is contributing its Model Context Protocol (MCP), which provides a standardized way to connect AI models and agents to data sources and tools. OpenAI is bringing AGENTS.md, a simple instruction file for developers that defines how AI coding tools should behave. Block, best known for its fintech products like Square and Cash App, is contributing Goose, its open-source agent framework already used by thousands of engineers for tasks like coding, data analysis, and documentation.
"Getting it out into the world gives us a place for other people to come help us make it better", said Block AI tech lead Brad Axen. He highlighted how open-sourcing Goose allows external contributors to refine the tool while benefiting Block internally: "We have a lot of contributors from open source, and everything they do to improve it comes back to our company."
sbb-itb-0f47b19
A Collaborative Push for Open Standards
The AAIF initiative is supported by other significant names, including AWS, Bloomberg, Cloudflare, and Google. This broad coalition underscores an industry-wide commitment to ensuring AI agents are trustworthy and interoperable at scale.
For OpenAI engineer Nick Cooper, the value of protocols lies in their ability to foster collaboration between systems. "We need multiple [protocols] to negotiate, communicate, and work together to deliver value for people, and that sort of openness and communication is why it’s not ever going to be one provider, one host, one company", Cooper said.
Jim Zemlin, executive director of the Linux Foundation, emphasized the importance of avoiding proprietary "closed wall" systems. "By bringing these projects together under the AAIF, we are now able to coordinate interoperability, safety patterns, and best practices specifically for AI agents", he said.
Open Source, Open Governance
The governance model of AAIF is designed to ensure no single vendor dominates the direction of its projects. Funding for the group comes through a "directed fund", with contributions from member companies. However, project roadmaps will be determined by technical steering committees, not by individual donors.
Anthropic’s MCP co-creator, David Soria Parra, outlined the goal of making the protocol a standard for connecting models, tools, and applications. "The main goal is to have enough adoption in the world that it’s the de facto standard. We’re all better off if we have an open integration center where you can build something once as a developer and use it across any client", he said.
Donating tools like MCP and Goose to the Linux Foundation demonstrates that these companies are prioritizing collaboration and neutrality over proprietary control. For Block, Goose also serves as a working example of how different tools, such as MCP and AGENTS.md, can integrate seamlessly.
A Vision for the Future
The AAIF’s ultimate ambition is to lay the groundwork for a mix-and-match ecosystem of AI tools, where agents operate within open, interoperable systems, much like the standards that underpin the modern web. Beyond the technical benefits of shared protocols, the initiative aims to establish safety patterns and best practices to ensure AI agents can be implemented responsibly.
Success, however, will require active adoption and evolution. "An early indicator of success, in addition to adoption of these standards, would be the development and implementation of shared standards being used by vendor agents around the world", said Zemlin.
Cooper echoed the importance of adaptability: "I don’t want it to be a stagnant thing. I don’t want these protocols to be part of this foundation, and that’s where they sat for two years. They should evolve and continually accept further input."
While the future of AAIF remains to be seen, its founders and contributors believe that open standards and collaboration will drive meaningful progress. By setting the stage for interoperability, they hope to avoid the pitfalls of proprietary silos and unlock new possibilities for developers and enterprises alike.