This Is How Trump Finally Signed the AI Executive Order
Original reporting by Wired
After nearly two weeks of internal deliberation and a previous shelving, President Donald Trump has signed a scaled-back executive order establishing a critical federal framework for artificial intelligence. This directive marks the first major AI regulation of his second term and underscores a growing concern within the administration: that increasingly powerful AI models, such as Anthropic’s Claude Mythos and OpenAI’s GPT-5.5, could be exploited to launch devastating cyberattacks against critical national infrastructure.
A strategic compromise The newly approved order grants the federal government exclusive access to the most advanced AI models 30 days before their public release. This significantly shorter review period—a departure from an earlier proposal of 90 days—was a direct result of intense negotiations and industry feedback, which had previously warned that a longer window would be unduly burdensome. White House Chief of Staff Susie Wiles, Treasury Secretary Scott Bessent, and National Cyber Director Sean Cairncross were pivotal in reviving the proposal after initial resistance, including from Trump’s former AI czar David Sacks. This compromise, which has garnered support from leading AI companies, allows officials vital time to proactively identify and address potential vulnerabilities in frontier models. With domestic policy now settled, the order also paves the way for the US to explore crucial international discussions, including with China, on creating similar cross-border frameworks for advanced AI systems, aiming to cement America's global leadership in balancing innovation with security.
The signing of President Trump's revised executive order marks a pivotal, if cautious, first stride in establishing a federal framework for artificial intelligence. By securing a 30-day window for government review of frontier AI models, the administration has initiated a process aimed at preempting potential national security threats, particularly those targeting critical infrastructure. This compromise, born from intense internal debate and industry negotiation, underscores the delicate balance policymakers face between fostering rapid innovation and mitigating the inherent risks of increasingly powerful AI systems. It also signals the administration's acknowledgment that outright delay in addressing these capabilities was no longer tenable, setting a precedent for engagement with the rapidly evolving sector and for navigating the complex interests of tech companies and national security agencies.
Shaping the AI landscape
Beyond its immediate operational directives, this executive order carries significant broader implications for the trajectory of AI development and governance. It establishes a collaborative, albeit voluntary, model that sets a clear expectation for future government-industry engagement, hinting at the potential for more formalized regulation as the technology matures and its societal impact becomes clearer. Crucially, the order enables the US to advance critical dialogues with nations like China, potentially leading to the development of much-needed cross-border frameworks essential for managing the shared risks of advanced AI on an international scale. This strategic move positions the United States not only as a leader in AI innovation but also as a key architect of its global oversight, reflecting a long-term commitment to balancing technological progress with robust security. The interplay between industry collaboration, national security imperatives, and international diplomacy will undoubtedly define the next phase of AI policy, with this order serving as a foundational blueprint for future advancements and challenges.