“Promoting Advanced AI Innovation and Security” Executive Order: Top points

On June 2, 2026, President Donald Trump signed an Executive Order (EO) titled, “Promoting Advanced Artificial Intelligence Innovation and Security.” 

The EO includes measures to address emerging cybersecurity risks presented by advanced artificial intelligence (AI) models, noting that these capabilities “introduce new national security considerations that require coordinated action across executive departments and agencies.” 

A new benchmarking process and voluntary framework for frontier AI models

The EO directs the Secretaries of the Treasury, War, and Homeland Security, acting through relevant components – including the National Security Agency (NSA) and the Cybersecurity and Infrastructure Security Agency (CISA) – and in consultation with other federal agencies to develop:

• A classified benchmarking process to assess the advanced cyber capabilities of AI models and determine the threshold at which they should be designated a “covered frontier model” for the purposes of the order
  
  
• A voluntary framework for AI developers to 1) engage with the federal government to determine whether models meet the threshold for designation and 2) grant agencies access to covered frontier models for 30 days prior to release to wider audiences

The framework would also include collaboration in selecting trusted partners for early access to these models. Notably, the EO explicitly states that it does not authorize any new mandatory government licensing, pre-clearance, or permitting requirements for the development, release, or distribution of AI models.  

The EO also includes provisions designed to strengthen cyber defenses for federal information systems, establish or expand cybersecurity support to government and critical infrastructure, and create new mechanisms for coordinated vulnerability management.

Further, the EO directs the Attorney General to prioritize enforcement of certain federal criminal laws applicable when AI is used to illegally access or damage a computer without authorization, or where AI is used while engaged in such illegal access to further any other crime (e.g., breaches of public or private information technology systems, or instances where AI agents are used to unlawfully access data that is subsequently used for criminal or unlawful purposes).

Broader context of the new EO

The EO comes amid increased attention from policymakers and the general public to the pace and scale of AI-enabled cyberattacks and vulnerability detection. It follows recent developments in which a select group of public- and private-sector professionals across the cybersecurity, technology, and critical infrastructure sectors were granted early access to a new AI product that could reportedly uncover software vulnerabilities at a pace beyond that of human researchers. 

Granting those in the defense space early access to models could support swift deployment of advanced AI capabilities for various cyber defense purposes, including scanning critical systems for vulnerabilities and installing patches to mitigate exploitation. Federal agencies may make similar use of unreleased models and leverage their unique authorities, capabilities, and insights to drive broader impact.  

Federal funding opportunity 

The EO also includes the following provision relevant to entities with advanced AI vulnerability detection capabilities: 

Within 30 days of the date of this order, the Director of OMB, in coordination with the National Cyber Director and the Director of CISA, shall determine whether any Federal grant programs have available and relevant funding that can be directed toward applicants developing advanced AI vulnerability detection.

Key takeaways

By signing this EO, the White House places increased emphasis on collaboration and coordination in addressing the current AI-enabled cyber threat environment. The order, which establishes a framework for ongoing public-private coordination and creates avenues to bolster the federal government’s response to new AI capabilities, relies on voluntary participation from AI, cybersecurity, and technology industry stakeholders. Industry willingness to engage with the public sector may be reflected in the number of AI companies with existing agreements with certain federal agencies – including the Department of War, NSA, and the Department of Commerce – to support collaboration and advanced model review. 

As EOs are not permanent, there have been calls for congressional action to codify standards and expectations of industry regarding advanced AI capabilities, including in areas such as cyber and Chemical, Biological, Radiological, Nuclear, and Explosives (CBRNE). While this EO pertains specifically to cybersecurity, similar mechanisms could emerge in CBRNE fields as frontier AI models continue to develop.

Finally, the EO may reflect a shift in federal roles and responsibilities under the Trump Administration. Historically, agencies such as CISA and the National Institute of Standards and Technology may have featured much more prominently in an EO focused on cybersecurity implications for advanced AI. If this reflects a broader trend, recent shifts in agency missions and operational focus areas could have longer-term impacts for agency capabilities, authorities, resourcing, and subject matter expertise. 

For more information, please contact the authors.