Tech Brief: AI stakes out The Hill

OpenAI has laid out a blueprint for a national AI regulation framework, signaling a shift toward formal federal standards. The plan builds on the existing executive order and state efforts, including California's SB 53, New York's RAISE Act, and Illinois' SB 315, while emphasizing independent risk evaluation, transparency and accountability. It also calls for strengthening the Center for AI Standards and Innovation into a permanent body under the secretary of commerce with national security oversight. Beyond governance, OpenAI argues for a resilience strategy that prioritizes AI within the U.S. tech stack while maintaining defensive safeguards.

OpenAI favors mandatory model evaluations but would limit CAISI's power to recommend vulnerabilities, leaving release decisions with the labs. The proposal contrasts with the executive order, which keeps disclosures voluntary and situates an evaluator within the intelligence community. Representative Seth Magaziner referenced unpublished draft legislation to place CAISI under Commerce. With the action stalled, the House plans to publish a discussion draft for a federal regulatory framework.