US Advances Sweeping AI Policy Overhaul With New Executive Order and Federal Legislation Draft

The first week of June 2026 marked a watershed moment in United States artificial intelligence policy. In a coordinated effort to centralize governance, the federal government advanced two major initiatives: a presidential Executive Order focused on national security, and a sweeping bipartisan legislative draft aimed at preempting state laws.[9]

The core of this overhaul represents a distinct pivot. Rather than focusing primarily on consumer safety or algorithmic bias, the new federal posture frames AI as a critical cybersecurity asset and a geopolitical arms race. The goal is to establish a unified national standard that prevents a fragmented patchwork of state regulations while maintaining American dominance in the sector.[1][9]

On June 2, President Trump signed the 'Promoting Advanced Artificial Intelligence Innovation and Security' Executive Order. The directive explicitly states a refusal to 'stifle this innovation with overly burdensome regulation,' marking a continuation of the administration's deregulatory stance on commercial AI.[1][2]

However, the order introduces aggressive new timelines for cyber defense. Within 60 days, the National Security Agency and the Cybersecurity and Infrastructure Security Agency must develop a classified benchmarking process to define what constitutes a 'covered frontier model.'[1][6]

Once that threshold is established, the government will ask developers to voluntarily share these frontier models for review up to 30 days before their public release. The stated objective is to allow federal agencies and critical infrastructure operators to identify software vulnerabilities and harden defenses before adversaries can exploit the new AI capabilities.[1][2][6]

While tech industry advocates successfully lobbied to keep this review process voluntary—avoiding rumored mandatory preclearance requirements—the order directs the Attorney General to prioritize prosecuting criminal actors who use AI to breach computer systems.[2][6]

Two days after the Executive Order, Representatives Jay Obernolte and Lori Trahan released a 269-page discussion draft titled the 'Great American AI Act.' The bipartisan proposal represents Congress's most comprehensive attempt to date to codify a national AI framework.[3][5]

The draft legislation's most contentious provision is a three-year preemption of state laws that regulate AI development. Lawmakers argue that advanced AI systems are built in one state and deployed nationwide, making zip-code-dependent protections untenable for developers.[3][5]

This federal push is a direct response to a rapidly fragmenting legal landscape. On January 1, 2026, multiple state laws took effect, including California's Transparency in Frontier AI Act—which mandates risk frameworks and safety reporting for massive models—and Texas's Responsible AI Governance Act. Colorado's comprehensive AI Act is also slated for implementation in mid-2026.[7]

If passed, the Great American AI Act would effectively freeze these state-level developer mandates. In their place, the bill would codify a Center for AI Standards and Innovation within the Commerce Department to oversee national evaluations and third-party audits of frontier models.[3][5]

While the preemption debate dominates headlines, the legislative draft introduces significant new obligations for employers regarding workforce displacement. The bill acknowledges that AI's most immediate societal impact may be economic rather than existential.[4][8]

The bill proposes amending the Worker Adjustment and Retraining Notification Act. Under the new rules, if an employer conducts a qualifying mass layoff where artificial intelligence was a 'substantial factor' in the decision or the displacement, the company must explicitly disclose that fact in its 60-day advance notice.[4]

Furthermore, the legislation would establish an AI Workforce Research Hub within the Department of Labor and require the Bureau of Labor Statistics to formally track AI adoption across the economy. This data collection is designed to inform future regulatory guidance and identify occupations most vulnerable to automation.[4][8]

Despite the momentum, the path forward remains highly uncertain. The Great American AI Act is currently only a discussion draft, and political analysts doubt it will pass before the August recess or the upcoming midterm elections.[4]

Additionally, the proposed federal preemption faces fierce opposition from civil society groups and state attorneys general, who argue it creates a regulatory vacuum by stripping local protections without imposing equally stringent federal mandates.[4][5]

As the August 2026 enforcement deadlines for the European Union's AI Act approach, the United States is racing to solidify its own alternative framework—one that relies heavily on public-private partnerships, national security imperatives, and a unified, business-friendly domestic market.[9]