EO 14409 — Advanced AI Innovation and Security
Status and scope
Section titled “Status and scope”Executive Order 14409, Promoting Advanced Artificial Intelligence Innovation and Security, was signed on 2 June 2026. It creates a national-security and cybersecurity coordination framework for advanced AI while expressly rejecting mandatory model licensing or preclearance.
Core measures
Section titled “Core measures”- Classified benchmarking: NSA, CISA, NIST and partner agencies must develop a classified process for assessing advanced cyber capabilities and determining when a model is a “covered frontier model.”
- Voluntary pre-release access: participating developers may provide the federal government access to a covered model for up to 30 days before release to other trusted partners.
- Critical-infrastructure defence: agencies are directed to expand AI-enabled cyber-defence tools and make relevant services available to federal, state and local bodies and critical-infrastructure operators.
- AI cybersecurity clearinghouse: Treasury, NSA, CISA and industry are directed to coordinate vulnerability discovery, validation, remediation and patch distribution.
- No licensing power: section 3(c) states that the order does not authorise mandatory licensing, permitting or governmental preclearance for model development or release.
Why it matters
Section titled “Why it matters”The order partially rebuilds a federal frontier-model interface after EO 14179 revoked the mandatory reporting structure created by EO 14110. The new system is narrower—centred on cyber capability and national security—and formally voluntary. It therefore does not restore the earlier 10²⁶-FLOP reporting rule.