In a flurry of activity in June 2026, U.S. AI policy advanced on multiple fronts, revealing a fragmented but fast-evolving regulatory landscape.

Federal Executive Action and Voluntary Model Reviews The Trump administration’s December 11, 2025 Executive Order 14365, titled Ensuring a National Policy Framework for Artificial Intelligence, continues to drive federal efforts to preempt state AI laws and establish a unified national standard (en.wikipedia.org). In March 2026, the White House released a seven-pillar AI policy framework urging Congress to legislate on issues ranging from children’s safety to intellectual property, while emphasizing preemption of conflicting state laws (nextgov.com).

Simultaneously, the administration has leaned on voluntary cooperation from AI developers. Major firms like OpenAI, Anthropic, Google, Microsoft, and xAI have agreed to submit frontier AI models for government review before public release (techradar.com). Meta is reportedly being urged to join this voluntary review program (techradar.com).

Anthropic’s Models: Export Controls Lifted After Security Review Anthropic’s advanced AI models—Fable and Mythos—were briefly blocked by export controls in June 2026 over national security concerns. After a two-week review, Commerce Secretary Howard Lutnick lifted the restrictions, citing Anthropic’s agreement to enhance security measures and collaborate on future standards (washingtonpost.com). This episode underscores the administration’s willingness to intervene directly in frontier AI deployment (washingtonpost.com).

State-Level AI Legislation Remains Active Despite federal preemption efforts, state AI laws continue to advance. As of May 2026, the U.S. lacks a comprehensive federal AI statute, leaving a patchwork of executive orders, agency guidance, and state laws in place (knowledgelib.io). Notable state statutes include California’s SB 53, New York’s RAISE Act, Illinois’s SB 315, and Texas’s TRAIGA, all imposing varying obligations and penalties on AI developers (knowledgelib.io). Colorado recently repealed its broader AI Act in favor of a narrower disclosure regime effective January 2027 (knowledgelib.io).

Congress Moves: Cloud Security Act Introduced On Capitol Hill, bipartisan lawmakers introduced the Cloud Security Act, which would require U.S. cloud providers to flag suspected foreign misuse of advanced AI computing to the Commerce Department (axios.com). The bill aims to close a loophole in chip export controls by extending oversight to cloud-based AI development—a growing vector for frontier model creation (axios.com).

Outlook: A Patchwork in Motion The U.S. AI policy environment in mid‑2026 is characterized by improvisation rather than comprehensive legislation. The administration’s reliance on executive orders and voluntary frameworks, combined with active state-level regulation and emerging congressional bills, has created a dynamic but fragmented governance regime (axios.com).

As Congress considers broader AI legislation, stakeholders must navigate overlapping mandates—from federal preemption efforts and voluntary model reviews to state-specific compliance regimes and new reporting obligations under proposed laws like the Cloud Security Act.