Trump Issues AI Cybersecurity Order, Convening Tech Giants on Frontier Model Review
Executive action establishes voluntary framework for government pre-deployment testing of advanced AI systems, marking first major governance move of second term.
President Trump is expected to sign an executive order on AI cybersecurity as soon as Thursday, establishing a voluntary framework requiring AI companies to submit advanced models for government review before public release—a sharp pivot from the administration’s hands-off approach to AI regulation.
The order focuses on two parallel tracks: strengthening Cybersecurity protections across federal agencies and critical infrastructure, and creating a pre-launch review process for frontier AI models. Under the voluntary agreement, companies would share advanced systems with the Commerce Department’s Center for AI Standards and Innovation for evaluation periods ranging from 14 to 90 days, according to CNN.
The move represents the first major AI governance action of Trump’s second term and signals a fundamental recalibration driven by National Security imperatives. The catalyst: Anthropic’s Mythos model, disclosed in April, which demonstrated unprecedented capability to identify zero-day vulnerabilities in widely-used software and infrastructure at machine speed.
The Commerce Department’s Center for AI Standards and Innovation has conducted 40 evaluations of frontier AI models through agreements with OpenAI, Anthropic, Google DeepMind, Microsoft, and xAI. However, the center operates with approximately 30 staff members and only $30 million in funding since its 2024 establishment—chronically underfunded for the scope of its mandate.
From Deregulation to Coordination
The executive order marks a significant shift in the Trump administration’s AI posture. Kevin Hassett, Director of the National Economic Council, framed the new approach as comparable to FDA drug approval. “We’re studying possibly an executive order to give a clear roadmap to everybody about how this is going to go and how future AI that also potentially create vulnerabilities should go through a process so that they’re released in the wild after they’ve been proven safe, just like an FDA drug,” he told reporters earlier this month, per Federal News Network.
The Mythos disclosure prompted urgent White House meetings with Anthropic leadership in April. Hassett described the government response: “We have scrambled an all-of-government effort and all the private sector to coordinate and to make sure that before this model is released out into the wild, that it’s been tested left and right to make sure that it doesn’t cause any harm to the American businesses or the American government.”
The voluntary framework reflects internal administration tensions between hardliners seeking mandatory controls and innovation advocates resisting regulatory burdens. Axios reported ongoing White House infighting over the order’s scope, with a senior official cautioning that “any Policy announcement will come directly from the president” and dismissing discussion of specific executive orders as “speculation.”
“We have scrambled an all-of-government effort and all the private sector to coordinate and to make sure that before this model is released out into the wild, that it’s been tested left and right to make sure that it doesn’t cause any harm to the American businesses or the American government.”
— Kevin Hassett, Director of the National Economic Council
Cybersecurity Component
Beyond frontier model review, the executive order aims to secure Pentagon and national security agencies while strengthening defenses at hospitals and banks. The cybersecurity provisions encourage threat intelligence sharing between AI companies and government, according to Axios. The order also seeks to boost cyber hiring across federal agencies.
This focus on critical infrastructure defense comes despite the administration’s earlier decision to cut funding and staffing at the Cybersecurity and Infrastructure Security Agency, creating tension between stated priorities and resource allocation.
- Voluntary framework gives government 14-90 day pre-launch review window for frontier AI models
- Order establishes public-private partnership model rather than mandatory regulation
- Cybersecurity provisions target federal agencies, Pentagon, hospitals, and financial institutions
- Center for AI Standards and Innovation operates with just 30 staff and $30 million funding
Industry Participation
The Commerce Department’s Center for AI Standards and Innovation has already established pre-deployment testing agreements with major AI developers including OpenAI, Anthropic, Google DeepMind, Microsoft, and xAI, CNBC reported earlier this month. These partnerships formed the foundation for the executive order’s voluntary framework.
The negotiated review timeframes—ranging from 14 days for incremental updates to 90 days for entirely new model architectures—reflect industry pushback against lengthy approval processes that could slow deployment cycles. Major tech companies conveyed concerns that extended review periods would disadvantage U.S. firms relative to international competitors operating without similar constraints.
Bloomberg reported that the administration deliberately omitted mandatory testing requirements, opting instead for voluntary partnership to maintain industry cooperation. This approach mirrors the administration’s broader preference for private-sector-led solutions over regulatory mandates.
What to Watch
The executive order’s effectiveness hinges on industry compliance with a voluntary framework that lacks enforcement mechanisms. If major AI developers bypass the review process—particularly for models developed outside U.S. jurisdiction—the administration may face pressure to implement mandatory requirements.
Resource constraints at the Center for AI Standards and Innovation present another challenge. With 30 staff members evaluating models from the world’s largest technology companies, throughput capacity could become a bottleneck. Congressional appropriations for CAISI expansion will signal whether lawmakers support the administration’s public-private partnership model.
The order also sets the stage for broader AI policy framework development. How the administration balances innovation incentives with national security imperatives in subsequent actions will determine whether this represents a permanent shift toward active AI governance or a temporary response to the Mythos disclosure. International coordination on frontier model evaluation standards remains conspicuously absent from the current framework, leaving open questions about cross-border AI deployment oversight.