President Donald Trump is poised to issue an executive order as soon as Thursday aimed at bolstering artificial intelligence cybersecurity. According to reports from Bloomberg News, the order would revamp existing cybersecurity information-sharing programs to include AI companies while stopping short of mandatory federal approval of cutting-edge models. Instead, it would call for voluntary government testing of frontier AI systems to find and patch weaknesses across federal, state and local networks, as well as critical US infrastructure, without requiring extensive new oversight.
The anticipated executive order includes a provision that would establish the National Security Agency as a key player in voluntary AI model testing. According to multiple sources familiar with the matter, the NSA is expected to handle classified testing of models offered up by AI labs before those models are publicly distributed. This represents a significant shift in intelligence community involvement in AI oversight, as the administration appears to prefer the intelligence community lead on AI model testing over the Commerce Department. An NSA spokesperson referred Nextgov/FCW to the White House, with a White House official stating that any policy announcement will come directly from the president.
Invitations have been sent to a range of technology industry executives for a signing event Thursday at the White House, though it's unclear who will take part. As reported by Bloomberg News, the directive is set to be signed about a month after Anthropic revealed that its breakthrough Mythos model was extraordinarily adept at finding network vulnerabilities and could pose a major cybersecurity risk. The company has limited Mythos access for now to a handful of large tech and Wall Street companies, amid broader global alarm about the new threats it could pose to critical systems. According to POLITICO, industry officials began receiving details on the directive from the White House on Tuesday night, with the order potentially coming as early as Thursday.
The draft executive order is expected to contain at least two sections focusing on cybersecurity and covered frontier models. As reported by POLITICO, the first segment gives the Pentagon 30 days to secure its networks, including key telecommunications and information systems, while the administration would have 30 days to mandate wider AI use across government systems and critical infrastructure organizations such as community banks, rural hospitals and utilities. The directive also tasks the Treasury Department with leading voluntary work with AI industry officials to form a clearinghouse within 30 days, establishing a voluntary partnership with AI-makers and critical infrastructure owners to find and patch vulnerabilities. Other federal agencies including the Office of the National Cyber Director, the National Security Agency and the Cybersecurity and Infrastructure Security Agency will support the effort.
The US already runs a voluntary program to evaluate AI systems before their release, and the Commerce Department recently announced an expansion of that initiative. According to Bloomberg News, Alphabet Inc.'s Google, Microsoft Corp. and xAI Inc. have agreed to give the government access to their models to assess the systems' capabilities and help improve security. OpenAI and Anthropic were already part of the program, led by the department's Center for AI Standards and Innovation. The second section of the new directive is expected to give a cohort of federal agencies and offices — including Treasury, CISA and the National Institute of Standards and Technology — 60 days to establish a classified benchmarking process to determine what constitutes a 'covered frontier model' under the executive order.
