On June 2, 2026, President Trump signed Executive Order (EO) 14409, titled “Promoting Advanced Artificial Intelligence Innovation and Security,” which established the administration’s latest policy related to artificial intelligence (AI). EO 14409 reiterates the administration’s stated commitment to fostering AI innovation and addresses both cyber- and national security concerns as AI capabilities become increasingly advanced.
The EO’s three primary objectives are: (1) strengthening the federal government’s cybersecurity protections, (2) ensuring that AI developers can collaborate with the federal government to enhance the security of their innovations, and (3) directing federal law enforcement to focus on criminal misuse of AI.
Upgrading American Systems for Advanced AI
Section 2 of the EO directs multiple federal agencies to take actions that strengthen federal information systems’ defense against cyber threats on an expedited timeline, with most initiatives requiring action by July 2, 2026. Specifically, the EO directs the following actions:
- The Committee on National Security Systems and the Secretary of Defense must prioritize the cyber defense of national security and military information systems.
- The Department of Homeland Security and the Cybersecurity and Infrastructure Security Agency (CISA) must issue guidance that strengthens cybersecurity of civilian federal systems, enhances AI-enabled defensive tools, and facilitates access to cybersecurity tools and services.
- The Secretary of the Treasury, among other agency leaders, must form an AI cybersecurity clearinghouse that scans for, validates, and remediates vulnerabilities.
- The Director of the Office of Management and Budget must identify federal grant programs with funding that could be directed toward applicants developing AI vulnerability detection technologies.
Securing Frontier Model Development
Section 3 of the EO provides for a framework that various agency leaders must develop by August 1, 2026, allowing the federal government and AI developers to collaborate on the most advanced AI systems, called “frontier models.” The focus of this section remains on enhancing security without creating regulatory burdens that could stymie AI innovation.
First, the EO directs federal agencies to develop a classified benchmarking process for evaluating the advanced cyber capabilities of AI models for the purpose of determining which constitute “covered frontier models.”
Second, once deemed a “covered frontier model,” the developers may voluntarily provide the federal government with access to the model for up to 30 days before it is released to customers, partners, or the public. During that period—subject to confidentiality, cybersecurity, insider-risk, and intellectual property protections—government experts and selected trusted partners can evaluate the model’s capabilities and cybersecurity strength.
Importantly, the EO expressly states that Section 3 does not create any licensing, permitting, or government approval requirements for developing or releasing AI models.
Protecting Against Criminal Actors
Section 4 of the EO directs the Department of Justice to prioritize the enforcement of existing federal criminal laws against anyone who misuses AI, including those who use AI to facilitate cyberattacks or access computer systems without authorization.
Conclusion
Although EO 14409 imposes no new mandates on the private sector, it appears to respond to public concern over AI risks and regulatory gaps. The EO tries to strike a balance by emphasizing the administration’s commitment to cybersecurity while sustaining its emphasis on AI innovation.