Trump administration seeks voluntary AI model cybersecurity tests in the U.S.

Traders Union
All News
Financial News
AI model cybersecurity tests

Highlights

President Trump signed an executive order directing Treasury, Defense, Commerce, and Homeland Security to secure voluntary cybersecurity testing agreements with AI developers.
U.S. agencies are granted up to 30 days to test AI models before their release beyond the government, potentially delaying launches and impacting developer profitability.
The order directs Treasury Secretary Scott Bessent to collaborate with AI firms and critical infrastructure providers, including banks, to detect and patch cybersecurity flaws in vital sectors.

Executive order expands federal AI security review

As reported by Reuters, U.S. President Donald Trump signs an executive order directing the Treasury, Defense, Commerce and Homeland Security departments, along with other agencies, to secure voluntary agreements with AI developers for cybersecurity testing of their most capable models.

Under the order, U.S. agencies get as long as 30 days to test models before they are released to organizations outside the government. The measure also tells agencies to put greater emphasis on strengthening cyber defenses across the federal government.

The move signals a shift in Trump's AI strategy toward a more active federal role in monitoring the technology's capabilities. That contrasts with his broader hands-off stance toward the tech sector since returning to office and his efforts to discourage state-level AI rules he opposes.

The voluntary testing framework could affect AI companies' profitability if it delays model launches or pushes developers to alter systems in response to security concerns. Trump had been expected to sign an AI executive order on May 21, but delayed it that day, saying he objected to some elements and did not want to weaken the U.S. position in AI competition with China.

Industry implications for critical infrastructure and developers

Anthropic, OpenAI and Alphabet's Google met with the U.S. government about cybersecurity, a senior U.S. official said in May during a call with reporters on the order's development. The companies do not immediately respond to Reuters requests for comment on the executive order.

Treasury Secretary Scott Bessent says the Treasury Department consulted banks while developing the order. The order asks Bessent to work with AI developers and critical infrastructure providers to scan software for cybersecurity flaws and develop patches, with critical infrastructure covering sectors vital to the U.S. economy such as banking, emergency services and hospitals.

Voluntary federal testing has already existed for several years, with companies including OpenAI and Anthropic submitting models for scrutiny by the Commerce Department's Center for AI Standards and Innovation, which operated under a different name during former President Joe Biden's administration. The department also said in May that Google, xAI and Microsoft had agreed to submit AI models for security testing, although those details later disappeared from its website.

Our earlier article covered the Trump administration’s plan to give U.S. agencies early access to frontier AI models for cybersecurity reviews ahead of public release. We explained that the order set a shortened 30-day review window—down from an earlier 90-day draft—and positioned the process as a compromise between security concerns and the push to avoid slowing major U.S. AI developers.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.