Key Points

• On June 2, 2026, President Trump issued an EO on AI cybersecurity and, days later, NSPM-11, signaling the administration's preference for voluntary engagement with industry partners while accelerating AI adoption across the national security enterprise.
• The EO reflects a lighter-touch approach than earlier drafts, shortening the government's pre-release access period for certain frontier AI models from 90 days to 30 days and expressly rejecting mandatory licensing or preclearance requirements.
• On June 4, 2026, Reps. Jay Obernolte (R-CA) and Lori Trahan (D-MA) released the Great American AI Act discussion draft, the most comprehensive bipartisan AI framework proposed to date, showing their intent to further refine the proposal based on stakeholder feedback.
• The proposal's temporary three-year preemption of state AI model development laws, which would expire unless reauthorized by Congress, has emerged as one of the most consequential and controversial aspects of the debate over a federal AI framework.

Introduction

On June 2, 2026, President Trump signed an Executive Order (EO), Promoting Advanced Artificial Intelligence Innovation and Security, directing a government-wide effort to strengthen cybersecurity capabilities through advanced artificial intelligence (AI) technologies while promoting continued U.S. leadership in AI innovation. The EO establishes a voluntary framework for engagement between frontier AI developers and the federal government, expands the use of advanced AI tools to support cybersecurity objectives, and signals the administration's preference for a light-touch approach to commercial AI regulation.

Additionally, the EO requires federal officials to issue “binding operational directives” to expand AI-enabled defensive tools across the government and critical infrastructure. Because these directives are mandatory, deadline-driven requirements for civilian agencies, not advisory guidance, they may significantly impact future contractor solicitations. In the short term, these directives will likely introduce new technical baseline requirements and accelerate procurement of AI-enabled cybersecurity tools, thereby creating near-term compliance and market implications for contractors and vendors supporting federal agencies and critical infrastructure.

For AI products that try to restrict military or weapons use through their own conditions of use, the EO does not override private contract terms by itself. But it does signal a federal policy push toward closer government-private collaboration on advanced AI for national security and cyber defense, which may make broad “no military use” terms harder to maintain in practice when defense customers expect access, testing or integration support. For vendors, the biggest near-term effect is likely on frontier-model governance, cybersecurity review and government access requests rather than direct combat-use permissions. If a provider’s conditions of use prohibit military applications, that restriction can still apply unless the vendor separately waives or revises it, but the executive order increases pressure to distinguish between prohibited offensive uses and permitted defensive, cyber or national-security uses.

The EO follows an initial draft order that was expected to be signed on May 21, 2026 but was delayed amid internal administration discussions regarding the appropriate balance between federal oversight of frontier AI models and continued AI innovation. The initial draft would have provided the federal government with up to 90 days of pre-public access to certain frontier models. The final EO shortens that period to up to 30 days and makes clear that it does not authorize mandatory licensing, preclearance or permitting requirements for the development, release or distribution of AI models. Days later, President Trump issued National Security Presidential Memorandum-11 (NSPM-11), reinforcing the administration's emphasis on accelerating AI adoption across the national security enterprise.

The changes from the initial draft, coupled with the EO's emphasis on voluntary engagement, suggest the administration intends to work with industry rather than impose broad regulatory requirements on frontier AI developers. In doing so, the administration appears to be prioritizing continued U.S. innovation and competitiveness while relying on collaboration with the private sector to address emerging cybersecurity and national security risks. NSPM-11 further demonstrates the administration's view that accelerating AI adoption is a strategic priority and should extend across defense and intelligence functions. Together, these actions are likely to shape ongoing congressional debates surrounding frontier AI governance, particularly as lawmakers consider how to balance innovation, oversight, national security and U.S. competitiveness.

A summary of the EO's agency directives and implementation timelines is outlined below.

Implementation Timelines

Within 30 Days (By July 2, 2026)

• The Committee on National Security Systems (CNSS) must prioritize the cyber defense of National Security Systems.
• The Department of Defense (DoD) must prioritize the cyber defense of DoD information systems.
• The Cybersecurity and Infrastructure Security Agency (CISA), in consultation with the Office of Management and Budget (OMB), the National Security Council (NSC) and the Office of the National Cyber Director (ONCD), must issue Binding Operational Directives and related guidance to strengthen civilian federal cybersecurity, expand AI-enabled defensive capabilities and facilitate access to cybersecurity services and, where appropriate, covered frontier models.
• The Department of the Treasury (Treasury), the National Security Agency (NSA), CISA and ONCD must establish an AI cybersecurity clearinghouse to coordinate vulnerability discovery, validation, remediation and patch distribution in voluntary collaboration with AI developers and critical infrastructure operators.
• OMB, ONCD, and CISA must determine whether existing federal grant programs can support applicants developing advanced AI vulnerability detection technologies.

Within 60 Days (By August 1, 2026)

• The Office of Personnel Management (OPM) must expand the U.S. Tech Force Information Cybersecurity Specialist hiring and placement pathways.
• Treasury, NSA, CISA, ONCD, the Assistant to the President for Science and Technology (APST) and the National Institute of Standards and Technology (NIST) must establish a classified benchmarking process to assess advanced AI cyber capabilities and determine when an AI system qualifies as a “covered frontier model.”
• Treasury, NSA, CISA, ONCD, APST and NIST must establish a voluntary frontier model engagement framework that would allow developers to seek federal input on covered frontier model designations, provide the government pre-release access to certain advanced AI models for up to 30 days and coordinate with the government on trusted early-access partnerships intended to improve critical infrastructure cybersecurity.

Ongoing

• The Department of Justice (DOJ) must pursue enforcement actions against individuals who use AI to facilitate unauthorized access to computer systems, cyber intrusions, data theft and related criminal activity.

House Introduces Discussion Draft of the Great American AI Act

On June 4, 2026, Reps. Jay Obernolte (R-CA) and Lori Trahan (D-MA), joined by Reps. Suhas Subramanyam (D-VA), Scott Franklin (R-FL), Scott Peters (D-CA) and Erin Houchin (R-IN), released a highly anticipated discussion draft of the Great American AI Act, which is the most comprehensive bipartisan AI regulatory legislative framework proposed to date. Released during a pivotal week for AI policy and after months of anticipation, the proposal seeks to establish a federal governance framework for increasingly capable AI systems that recognizes their implications for U.S. economic competitiveness and national security. Rather than formally introducing legislation, however, the members elected to release a discussion draft and invite feedback, signaling that lawmakers intend to further refine the proposal before introduction.

The discussion draft reflects an effort to establish a more durable, uniform national approach to frontier AI model regulation. In an accompanying op-ed, Reps. Obernolte and Trahan argued that President Trump's EO is insufficient and that lasting rules for frontier AI systems will ultimately require congressional action. The lawmakers also presented the proposal as an attempt to balance innovation with appropriate safeguards while ensuring that the United States remains globally competitive in AI. Members of the public have been invited to submit comments to GAAIA@mail.house.gov.

Debate Over Federal Preemption

Perhaps the most consequential and controversial provision in the discussion draft is Section 121, which would preempt state and local laws specifically regulating AI model development for three years while preserving generally applicable laws and those governing the use and deployment of AI systems. The provision reflects an ongoing debate over who should be responsible for regulating AI: states, the federal government, or some combination of the two.

Proponents of federal preemption argue that a national framework would provide greater certainty for developers and users and avoid a patchwork of potentially conflicting state requirements. Reps. Obernolte and Trahan argue that AI risks do not stop at state lines and that protections should not depend on geography, emphasizing that the most advanced AI systems are developed in one state and deployed nationwide.

The proposal has already drawn mixed reactions from both sides of the aisle. In a statement following the release of the discussion draft, the House Democratic Caucus Commission on AI and the Innovation Economy stated that it does not support the proposal in its current form and questioned whether it could serve as the basis for productive dialogue. House Majority Leader Steve Scalise (R-LA) has expressed similar concerns that the framework's protections for innovation may be too limited, while also signaling support for broader federal preemption than that contemplated by the discussion draft.

Given the increasing pace of state AI legislation and the lack of consensus within Congress, the scope and duration of any federal preemption provision are likely to remain key issues as lawmakers continue to refine the proposal. Moreover, the limited time remaining in the 119th Congress and the absence of a comparable Senate proposal present additional hurdles to advancing comprehensive AI legislation. While Reps. Obernolte and Trahan have emphasized that the discussion draft is intended to facilitate stakeholder feedback and continued bipartisan discussions, the mixed reactions to the framework underscore the challenges lawmakers face in forging consensus around a federal approach to AI regulation.

The discussion draft is organized into four titles:

• Title I — Frontier Artificial Intelligence Governance: Establishes a federal oversight framework for advanced AI systems centered on new transparency, reporting and risk management requirements for large developers. The requirements apply primarily to “large frontier developers,” defined as companies that develop frontier AI models and exceed $500 million in annual revenue, as well as other developers of frontier models at the point of deployment. Covered companies would be required to publish risk frameworks, disclose model capabilities and mitigation measures and report safety incidents to regulators. The title also formally codifies the Center for AI Standards and Innovation (CAISI), empowering it to develop standards and oversee third-party audits, requires independent verification of compliance and provides whistleblower protections. Most notably, the title would temporarily preempt state and local laws specifically regulating AI model development for three years while preserving generally applicable laws and those governing AI use and deployment.
• Title II — Workforce: Focuses on preparing the labor market for AI adoption through expanded federal investment in education, skills development and workforce training, alongside a new federal data and research infrastructure on AI’s economic effects. The title directs the Department of Labor (DOL) to improve data collection, forecasting and analysis of AI-driven workforce changes and requires new employer disclosures where AI contributes to mass layoffs. It also lays the groundwork for potential worker adjustment assistance programs and requires integration of AI-driven labor market insights into federal and state workforce policy.
• Title III — Cybersecurity: Strengthens cybersecurity policy as it relates to AI by extending federal information-sharing authorities, supporting the security of critical open-source software and directing further study of vulnerabilities tied to AI systems and infrastructure. The title also seeks to improve coordination between government and industry on emerging risks associated with AI models, data centers and software supply chains.
• Title IV — Research, Development and International Cooperation: Advances U.S. leadership in AI by expanding federal research infrastructure, testing capabilities and public-private collaboration. The title establishes AI testbeds, supports access to data and compute resources and promotes the development of technical standards through international partnerships. It also emphasizes coordination with allied countries, investment in innovation and safeguards around research security.

A section-by-section breakdown of the discussion draft can be found here.

Conclusion

The administration's recent actions, along with the release of the Great American AI Act discussion draft, reflect a growing belief among policymakers that increasingly capable AI systems present both opportunities and risks that will require sustained policy attention. While disagreements remain over the appropriate degree of oversight, the role of the federal government relative to the states and the mechanisms through which AI should be governed, policymakers appear increasingly aligned on the need to develop a framework capable of addressing rapidly advancing AI systems.

At the same time, both the administration and Congress have signaled a willingness to engage closely with outside stakeholders as they navigate what remains largely uncharted territory. The administration's emphasis on voluntary engagement with frontier AI developers and congressmembers’ decision to release a discussion draft and solicit public feedback suggest policymakers recognize that effective AI governance will likely require continued collaboration with industry and the broader public. For companies, these developments present an opportunity to help shape the policies and standards that may ultimately govern increasingly powerful AI systems.

Although the path toward a comprehensive federal framework remains uncertain and may ultimately preserve a significant role for state-level regulation, President Trump's EO, NSPM-11 and the release of the Great American AI Act discussion draft suggest that the United States is entering a new phase of AI governance. In that environment, federal procurement, agency mandates and voluntary engagement frameworks are likely to operate together as complementary tools shaping the development and deployment of AI technologies. Akin's team will continue to closely monitor developments across the Executive Branch and Congress and keep clients apprised of noteworthy legislative and regulatory activity.