The Executive Order "Promoting Advanced Artificial Intelligence Innovation and Security" was signed by President Donald J. Trump on June 2, 2026. It is a cybersecurity-and-frontier-security order rather than a broad AI-policy order: it directs AI-enabled federal cyber defense on mostly 30-day clocks, establishes a Treasury-led voluntary AI cybersecurity clearinghouse, creates a classified benchmarking process for designating a "covered frontier model" with the threshold determination made by the NSA Director, and sets up a voluntary framework offering the government up-to-30-day pre-release access to such models — while explicitly barring any mandatory licensing or preclearance requirement. It is built around government visibility rather than government approval.
The signature block reads "DONALD J. TRUMP, THE WHITE HOUSE, June 2, 2026." This page summarizes the signed primary text retrieved from the official White House presidential-actions page. For the unsigned predecessor and its postponement story, see the superseded Draft Executive Order: Promoting Advanced Artificial Intelligence Innovation and Security (unsigned, May 2026); for the live tracking page, see EO — Promoting Advanced AI Innovation and Security (Trump, signed June 2, 2026).
Primary text: whitehouse.gov
Framing
Section 1 frames AI as making the United States "stronger" while introducing "new national security considerations," and pairs an "America First cybersecurity" posture with continued "global AI dominance." It front-loads a deregulatory thesis, stating that the U.S. leads "because we refuse to stifle this innovation with overly burdensome regulation," and credits the administration with "slashing the bureaucratic constraints that the prior administration placed on America's AI developers."
Key provisions
Section 2, "Upgrading American Systems for Advanced AI," runs on mostly 30-day clocks. The Committee on National Security Systems is directed to prioritize cyber defense of National Security Systems, and the Secretary of War to prioritize Department of War information-system cyber defense (the order uses the renamed "Department of War"). DHS, through CISA — with OMB, the APNSA, and the National Cyber Director — is to issue Binding Operational Directives to (i) expedite cyber defense of civilian federal systems, (ii) expand AI-enabled defensive cybersecurity programs and services, and (iii) facilitate access to cybersecurity tools "including where appropriate covered frontier models" for agencies, state and local authorities, and critical-infrastructure operators such as rural hospitals, community banks, and local utilities. Treasury — with the National Cyber Director, NSA (under the Secretary of War), and CISA — is to form an AI cybersecurity clearinghouse, in voluntary collaboration with industry and critical-infrastructure operators, to coordinate vulnerability scanning, validation, and patch distribution. OMB (with ONCD and CISA) is to determine within 30 days whether any federal grant programs can fund advanced AI vulnerability detection, and OPM, within 60 days, is to expand U.S. Tech Force cybersecurity hiring pathways.
Section 3, "Secure Frontier Model Deployment," runs on a 60-day clock. Treasury, NSA (under the Secretary of War), and CISA — with the National Cyber Director, the APST, and Commerce through NIST — are directed to: (a) develop and maintain a classified benchmarking process to assess models' advanced cyber capabilities and set the threshold for designating a "covered frontier model," with the NSA Director making the designation in consultation with ONCD, the APST, CISA, and Department of War representatives; (b) design a voluntary framework under which developers can ask the government whether a model meets "covered frontier model" status, provide the government access to covered frontier models for up to 30 days before release (subject to confidentiality, cybersecurity, insider-risk, and IP protections), and collaborate on selecting "trusted partners" for early access; and (c) explicitly bar construing the section to authorize "a mandatory governmental licensing, preclearance, or permitting requirement for the development, publication, release, or distribution of new AI models, including frontier models."
Section 4, "Protection Against Criminal Actors," directs the Attorney General to prioritize enforcement of 18 U.S.C. §§ 1028, 1030, 1343 and other criminal laws against AI-enabled unauthorized computer access and AI-agent-assisted crime.
Section 5, "General Provisions," contains standard savings clauses: no impairment of existing authority or OMB functions; subject to law and appropriations; and creating no enforceable right or benefit. Publication costs are borne by the Department of War.
Changes from the draft
The signed order is scaled back from the May draft whose progress was interrupted by David Sacks's May 21 phone call. The pre-release government access window was cut from up to 90 days in the draft to up to 30 days in the signed text, a change that addresses the friction point in the May postponement (the window industry sources called "too long").
| Provision | Draft (postponed May 21) | Signed (June 2) |
|---|---|---|
| Pre-release government access window | up to 90 days | up to 30 days |
| Licensing/preclearance | barred (Sec. 3(c)) | barred and reinforced (Sec. 3(c)) |
| "Covered frontier model" designation | classified benchmarking process | classified process; NSA Director designates (explicit) |
| Commerce/NIST role | implied | explicit ("through the Director of NIST") |
| Treasury clearinghouse lead | Treasury + NSA + CISA | retained (Treasury lead survived May criticism) |
Relation to other AI-security measures
The order institutionalizes limited government oversight of frontier models while foreclosing the mandatory pre-release ODNI-led review that a 32-member bipartisan House letter and an ICBA-led industry coalition both rejected in May 2026. It was positioned by the administration as a response to the frontier-model cyber-capability increase associated with the Mythos preview and GPT-5.5-Cyber. It advances information-sharing and frontier-cyber goals from America's AI Action Plan through executive authority rather than legislation.
The order was signed the same day Senator Gillibrand introduced the Secure and Accountable Military AI Act; the two June 2 actions represent contrasting approaches, voluntary visibility versus statutory guardrails.
Unlike the EU AI Act and the GPAI Code of Practice, the order imposes no binding obligations on developers, since access is opt-in. Unlike state transparency laws such as SB 53, the NY RAISE Act, and Illinois SB 315, it targets cyber capability and pre-release security rather than public disclosure of safety policies. Its "covered frontier model" threshold is set by a classified process rather than by statute.
Relationships
- supersedes: Draft Executive Order: Promoting Advanced Artificial Intelligence Innovation and Security (unsigned, May 2026) — signed text replaces the postponed predecisional draft as the canonical primary source.
- instance-of: AI and Cybersecurity · AI Pre-Release Vetting
- contradicts: mandatory frontier-model licensing / ODNI-led pre-release review proposals; the guardrails-first Secure and Accountable Military AI Act (Gillibrand, June 2026) (same-day contrast).
- related: EO — Promoting Advanced AI Innovation and Security (Trump, signed June 2, 2026) · America's AI Action Plan · Claude Mythos Preview · Dual-Use Frontier AI · Compute Governance · Cybersecurity and Infrastructure Security Agency (CISA) — regulator role · Donald Trump · David Sacks