Helping build shared standards for advanced AI

As increasingly capable artificial intelligence models offer benefits alongside potential safety and security risks, societies require robust institutions to evaluate and govern these systems. To address this need, OpenAI helped found the Appia Foundation, which is hosted by the Linux Foundation. Appia will develop open, modular specifications designed to translate established frameworks and international standards into practical assessment criteria across the AI value chain. This work is intended to establish a critical trust layer enabling third parties to check conformity and produce reusable evidence when models and applications are developed by different organizations. According to OpenAI, this effort builds upon a broader body of work that includes a blueprint for democratic governance, strengthened national institutions like the U.S. Center for AI Standards and Innovation and the UK AI Safety Institute, and transparent third-party evaluations. OpenAI utilizes its own Preparedness Framework and Frontier Governance Framework to manage serious risks, define internal practices, and address regulatory obligations. Appia's mission represents a subsequent challenge of making these practices interoperable across organizations, supply chains, and jurisdictions. Through these collective forums, the stated goal is to translate lessons learned from frontier development into open, technically grounded practices that companies, governments, and independent assessors can utilize globally.