The Trump administration has drafted an executive order that would create a framework for AI developers to voluntarily share their most advanced models with federal agencies before releasing them to the public. The key word there: voluntarily.
Under the proposed framework, developers of frontier AI models could provide access to their technologies up to 90 days before public release, giving government agencies time to conduct security reviews. No company would be required to participate. The order explicitly prohibits any mandatory government review or licensing requirements.
What the order actually does
The draft executive order tasks several federal agencies with building out the voluntary oversight process. The Cybersecurity and Infrastructure Security Agency (CISA), the Office of the National Cyber Director, and the Office of Science and Technology Policy (OSTP) would collaborate on identifying which AI models might benefit from a pre-release federal review. The NSA could also play a role.
The framework builds on Executive Order 14179, signed on January 23, 2025, which aimed to remove barriers to US AI leadership.
The initial signing of this new executive order was scheduled for May 21, 2025, but was postponed. President Trump reportedly expressed concerns that the order could negatively impact US AI competitiveness, even with its voluntary structure.
A deliberate departure from prior AI policy
This approach represents a significant shift from previous administrations’ handling of AI governance. Where earlier frameworks leaned toward establishing guardrails, compliance requirements, and structured oversight mechanisms, the Trump administration has consistently favored a light-touch regulatory approach.
The 90-day pre-release window is notable, though. Even in a voluntary context, three months of government access to a frontier model before it reaches the public is a meaningful amount of time. If major AI labs participate, federal agencies could identify potential security vulnerabilities, dual-use capabilities, or national security risks well before these systems are widely deployed.
What this means for the AI and crypto markets
For investors in the broader tech and AI space, the voluntary framework sends a clear signal: the current administration is not coming for your AI projects with a regulatory sledgehammer. Companies developing frontier models, whether in pure AI, decentralized AI networks, or AI-integrated crypto protocols, can operate with reasonable confidence that compliance costs won’t spike in the near term.
That’s particularly relevant for the growing intersection of AI and crypto. A mandatory federal review process for frontier models could have created significant complications for decentralized projects where no single entity controls model deployment. The voluntary approach sidesteps that problem entirely.
The postponement of the signing also introduces a layer of unpredictability. If the administration’s own voluntary framework was deemed too burdensome for competitiveness, it raises the question of whether any formal AI policy will emerge during this term.