Introduction to the Executive Order
The Trump administration released a revised executive order on artificial intelligence (AI) on Tuesday, offering a significantly pared-back vision for the federal government’s role in vetting AI systems. This order keeps in place the administration’s largely voluntary framework for companies to engage with the federal government around testing new models before release.
The order appears to considerably weaken or loosen provisions that had been opposed by industry. Under the order, AI companies would voluntarily provide the federal government access to frontier models before release, but now it will be for “up to” 30 days instead of the 90-day timeline included in previous drafts.
Key Provisions of the Order
The order explicitly states that nothing in the program will be construed as mandatory or part of a federal licensing or permitting regime, and gives AI companies significant influence to help define what models would and would not be covered under for testing.
It also states that all federal testing and access to the models would be subject to “confidentiality, cybersecurity, insider-risk, and intellectual-property protection, use, and nondisclosure requirements.”
Industry Reaction and National Security Implications
The White House increasingly sees national security implications in the rapid release of frontier models from the private sector, but it has also been one of the loudest critics of regulating the technology for fear it could harm American businesses.
“The United States continues to lead the world in Artificial Intelligence (AI) because of the enormous talent and innovation of our AI industry, and because we refuse to stifle this innovation with overly burdensome regulation,”
the order reads. This argument was bolstered in recent days as industry members and top advisers to Trump, like tech investor and AI czar David Sacks, lobbied against previous draft language, arguing it would put too much of a regulatory burden on U.S. businesses.
Implementation and Politicization Concerns
The order puts the Department of Treasury at the head of a new interagency cybersecurity clearinghouse on AI, where the private sector, critical infrastructure operators, and federal agencies voluntarily collaborate to coordinate and deconflict scanning for software vulnerabilities, discovery and validation and remediation activities, like patching.
Treasury, the Cybersecurity and Infrastructure Security Agency, the NSA, the Office of the National Cyber Director, and other agencies would also be responsible for developing classified benchmarks that would be used to identify or flag the kind of advanced cyber and hacking capabilities that agencies are interested in testing.
Questions linger over implementation, with some experts expressing concerns about the potential for politicization. Senator Mark Warner, D-Va., said the order would help the White House “begin to grapple” with the threats that new frontier models and their hacking capabilities pose to critical infrastructure, but was also sharply critical of the administration’s about-face on the need for federal scrutiny of emerging AI technologies.
Expert Reaction and Future Directions
Samir Jain of the Center for Democracy and Technology said that while AI models pose real cybersecurity threats to critical services, the order “attempts to avoid the deeply concerning implications of a mandatory licensing regime for release of new models.”
“Testing and benchmarking programs are important to promote cybersecurity and address other risks,” Jain said in a statement. “However, the EO should not become a mechanism for the Administration to punish companies for political or other arbitrary reasons, and so we will be closely monitoring the details of its implementation as they emerge.”
The full order can be read on the White House’s website.
Source: CyberScoop