Trump shifts US national security AI policy toward faster adoption

In short: The Trump administration is pushing US defense and intelligence agencies to use advanced AI faster, while keeping a narrower set of security checks.

What happened

The White House has changed direction on how the US government handles AI for national security. Compared with the Biden approach in 2024, the new focus puts more weight on rapid adoption, close work with AI companies, and keeping an edge over China.

In June 2026, President Trump signed a policy called NSPM-11. It tells national security agencies to speed up AI use across their missions and to reduce barriers that slow deployment. It also calls for deeper partnerships with industry so the most advanced models can be available to national security users “without delay.”

The administration also issued a separate AI order in June 2026 that emphasizes protecting US intellectual property (ideas and inventions that are legally protected) and strengthening systems against outside attacks. By July 2026, the White House said it would set up an AI and cybersecurity coordination group to help AI developers and critical infrastructure providers share information about weaknesses and respond faster.

At the same time, the government is not removing all guardrails. The White House has discussed having AI developers voluntarily submit their most capable “frontier” models, meaning the newest and most powerful systems, for cybersecurity testing before a public release. Agencies are also being told to define which frontier models would fall under these checks.

Why it matters

AI tools are starting to shape military planning, intelligence work, and cyber defense. A faster rollout could mean quicker gains for the US, but it can also raise the risk of mistakes or security gaps if checks do not keep up. For most people, the impact shows up indirectly, through how safe critical services are, like power, banking, and communications, and how the US manages competition with China.

Source: NYTimes