Overview
- The White House signed Executive Order 14409 on June 2, 2026, creating a voluntary framework for government review of advanced AI models before they go public.
- Agencies have 60 days to define what counts as a “covered frontier” model, a decision that will determine which labs and models face review and whether the policy targets a few large developers or a broader set.
- Under the order, developers may give government teams up to 30 days of pre-release access for cybersecurity testing and remediation, with industry still responsible for securing its own models.
- The Treasury will run a cybersecurity clearinghouse to share vulnerability data and coordinate fixes, and agencies will stand up classified benchmarking and defensive priorities led by national security partners.
- Reports say the administration asked OpenAI to limit GPT-5.6 rollouts to vetted partners while evaluations proceed, and experts warn the voluntary approach could create a two-tier market that squeezes startups, crypto projects, and GovTech that rely on open model access.