OpenAI Blueprint for Democratic Governance of Frontier AI

OpenAI has proposed a blueprint for the United States to establish a durable federal framework for governing frontier AI. This proposal aims to create a governance structure that can evolve alongside the rapid advancement of AI capabilities to ensure public safety and national security.

Three-Part Strategy for Federal AI Governance

The OpenAI blueprint outlines a three-pronged approach to building a national framework for AI safety:

1. National Framework Based on State Consensus

OpenAI proposes building a national framework that leverages the emerging consensus seen in various state-level frontier safety laws. The blueprint specifically references the following legislative efforts as foundations for federal action:

• California's SB 53
• New York's RAISE Act
• Illinois's SB 315

2. Strengthening CAISI

The blueprint identifies the Center for AI Safety and Innovation (CAISI) as the primary federal institution for frontier AI safety. OpenAI advocates for strengthening this institution to serve as the central hub for safety oversight and technical expertise within the U.S. government.

3. Government-Wide Resilience Plan

To address the national security and public safety challenges posed by frontier AI, OpenAI proposes the mobilization of a broader resilience plan across the entire federal government. This plan is intended to ensure that government agencies are prepared for the systemic risks associated with highly capable AI systems.

Context and Legislative Landscape

The proposal comes at a time of increasing regulatory activity at both the state and federal levels. OpenAI notes that the White House's executive order on "Promoting Advanced Artificial Intelligence Innovation and Security" is a critical step forward in establishing a baseline for AI security.

By integrating state-level harmonized approaches with federal oversight, OpenAI argues that the U.S. can build a governance system that is durable and capable of adapting to the technological shifts inherent in frontier AI development.