Helping build shared standards for advanced AI

Increasingly capable models can strengthen cyber defense, accelerate scientific discovery, and expand access to expertise. But they can also create safety and security risks if their capabilities are misunderstood, their safeguards are inadequate, or governments lack the information they need to respond. To realize the benefits safely and confidently, societies will need institutions with the technical and governance capacity to evaluate, secure, and govern increasingly capable systems.

That is one reason OpenAI helped found the Appia Foundation⁠(opens in a new window), hosted by the Linux Foundation. Appia will develop open, modular specifications intended to translate international standards and established frameworks into practical assessment criteria across the AI value chain. Its work can help develop a critical missing trust layer by which third parties check conformity with standards, producing clearer and more reusable evidence when models, infrastructure, and applications are developed by different organizations. In doing this work, Appia will help create a shared technical language that will allow national and international institutions to trust each other’s work.

We see this effort as an important next step in a broader body of work to strengthen the institutions, standards, and assessment practices needed for advanced AI systems.

Our recent blueprint for democratic governance of frontier AI offers a roadmap for that work. It calls for a durable U.S. framework, a strengthened Center for AI Standards and Innovation (CAISI), and a broader resilience strategy across government. It also recognizes that frontier risks are international in scope. Nations should work together to develop compatible safety frameworks, trusted channels for sharing risk findings, and coordinated responses to incidents.

Appia’s work is aimed at the next challenge: making those practices interoperable across organizations, jurisdictions, and the supply chain.