From the journal

U.S. Executive Order Establishes Voluntary Frontier AI Pre-Release Review, 2 June 2026

On 2 June 2026, President Trump signed the Executive Order 'Promoting Advanced Artificial Intelligence Innovation and Security,' directing federal agencies to develop benchmarks for assessing AI models' cyber capabilities, establish an AI cybersecurity clearinghouse, and strengthen government defenses against AI-enabled threats. The Order creates a voluntary program under which AI developers may submit frontier models for government review up to 30 days before releasing them to other trusted partners. No mandatory licensing, pre-clearance, or permitting requirement is created.

2 min read

President Trump signed the Executive Order 'Promoting Advanced Artificial Intelligence Innovation and Security' on 2 June 2026. The Order is effective upon signature. It imposes no mandatory regulatory burden on AI developers and directs federal agencies to build AI cybersecurity infrastructure.

The Order rests on the President's constitutional authority and does not amend or repeal any statute. Three directives govern agency conduct. First, agencies must develop benchmarks to assess AI models' capacity to accelerate offensive cyber operations. Second, the executive branch must create an 'AI cybersecurity clearinghouse' to collect, review, and share information on model vulnerabilities. Third, agencies must deploy AI tools to harden government networks and critical infrastructure. The General Provisions section states that nothing in the Order authorises any mandatory governmental licensing or pre-clearance requirement for the development, publication, or distribution of AI models.

AI developers building frontier models may submit pre-release versions to federal evaluators up to 30 days before distributing to other trusted partners. Participation is voluntary; no regulatory consequence attaches to firms that decline. This voluntary window replaced a previously circulated draft that proposed a 90-day mandatory review. The clearinghouse mechanism creates government visibility into AI capability disclosures made voluntarily by participating companies.

The Order does not define 'frontier AI model' with numerical thresholds. Agency implementing guidance must determine which systems fall within scope of the voluntary review. The absence of mandatory review distinguishes this Order from some allied-nation approaches and from earlier U.S. draft proposals. Whether future legislation or agency rulemaking will impose binding pre-release requirements is an open question. Implementation timelines for the cybersecurity clearinghouse and benchmark development are subject to agency discretion.

Licentium advises AI developers and deployers on U.S. federal AI policy, voluntary review programs, and cybersecurity obligations affecting AI products. We may assist directly or through our partner network on regulatory engagement with federal agencies. Work we undertake includes AI regulatory strategy, frontier model compliance programs, and AI-enabled cybersecurity obligations under federal procurement and critical-infrastructure rules.

Source: Executive Order 'Promoting Advanced Artificial Intelligence Innovation and Security,' White House, 2 June 2026

AI Regulatory

More from the journal

See all

MAS and Industry Publish AI Risk Management Toolkit for Singapore Financial Sector, 2026

The Monetary Authority of Singapore concluded Project MindForge Phase 2 in early 2026, publishing an AI Risk Management Operationalisation Handbook developed with a consortium of 24 banks, insurers, and capital market firms. The handbook provides practical implementation guidance across traditional AI, generative AI, and agentic AI systems, and applies alongside the MAS Guidelines for Artificial Intelligence Risk Management to establish supervisory expectations for Singapore-regulated financial institutions.

House of Lords Committee Publishes Report on UK Stablecoin Regulation, 3 June 2026

On 3 June 2026, the House of Lords Financial Services Regulation Committee published 'Stablecoins: waiting for regulation,' assessing the Bank of England's and the Financial Conduct Authority's proposed regulatory regimes for stablecoins in the UK. The Committee broadly supports the proposals but recommends reconsideration of holding limits, the requirement for unremunerated backing assets, and the proposed restriction on commercial banks issuing fiat-backed stablecoins.

FCA and Bank of England Call for Input on UK Wholesale Market Tokenisation, May 2026

On 18 May 2026, the Financial Conduct Authority, the Bank of England, and the Prudential Regulation Authority published a joint call for input setting out a shared vision for the safe adoption of tokenisation in UK wholesale financial markets. The consultation covers tokenised bonds, equities, and fund units and closes 3 July 2026. Responses will inform a joint roadmap aligned with the Government's Wholesale Financial Markets Digital Strategy.

Ready to launch without the regulatory guesswork?

Book a 30-minute consultation. We'll map your AI or licensing path and tell you exactly what's required, in plain language.