US Prepares AI Security Order Without Mandatory Model Tests

The Trump administration is preparing an executive order to protect U.S. networks from AI-enabled cyber attacks through partnerships with artificial intelligence companies [1, 2].

This move signals a shift in how the federal government balances national security with the pace of technological innovation. By avoiding mandatory requirements for AI developers, the administration is prioritizing industry collaboration over rigid regulatory oversight.

The proposed order directs U.S. agencies to work alongside AI firms to bolster defenses against emerging threats [1, 2]. The primary goal is to address security concerns as AI tools become more capable of automating and executing complex cyber attacks [1, 2].

Notably, the draft omits requirements for mandatory pre-release reviews of AI models [1, 2]. This means companies would not be forced to submit their systems for government testing or safety audits before making them available to the public, a departure from some previously discussed safety frameworks.

The administration is focusing on a cooperative model where agencies and private firms share intelligence to guard critical infrastructure [1, 2]. This approach seeks to mitigate the risks of AI-driven vulnerabilities without slowing the deployment of new tools.

Government officials have not yet released a final timeline for the order's implementation. The focus remains on creating a flexible security posture that can evolve as AI capabilities change [1, 2].