OpenAI Publishes Federal AI Regulation Blueprint
OpenAI, a leading AI developer, is publicly advocating for federal oversight of its own most advanced systems, arguing that company-only release decisions are insufficient and that state-level rules cannot address cross-border risks.
Reporting from 1 sources: GIGAZINE.
OpenAI released a policy proposal titled 'Democratic Governance of Frontier AI: A blueprint for a federal framework,' calling for a federal-level system to evaluate and manage high-performance frontier AI. The proposal centers on safety evaluations, a specialized evaluation system under the Center for AI Standards and Innovation (CAISI), and strengthened government defenses in cybersecurity and public health.
The proposal, published June 4, argues that decisions on releasing frontier AI should not be left solely to companies. OpenAI calls for a federal framework based on democratic procedures, warning that different rules across states would complicate corporate responses and fail to address risks that cross national and state borders.
Three main points anchor the blueprint: a national system requiring safety evaluations and transparency reports for frontier AI, a specialized evaluation system centered on the Center for AI Standards and Innovation (CAISI), and a strategy to strengthen government defenses in cybersecurity and public health. Specific measures include risk assessment of serious risks, independent audits, reporting of major incidents, protection of unreleased model weights, and whistleblower protection.
OpenAI states that while the most advanced AI should undergo pre-release evaluation, government agencies should not act as gatekeepers that permit or prohibit release. The company envisions a system where evaluation results and safety measures are made public, with the developing company bearing final responsibility. The proposal also notes that AI capabilities may advance faster than existing systems and rules can keep up, calling for early establishment of evaluation systems before frontier AI significantly impacts the economy, scientific research, or national security.
Synthesized by Yomimono from the 1 cited source below, including Japanese-language reporting where cited, then editorially reviewed before publishing.