U.S. President Donald Trump signed an executive order requiring artificial intelligence (AI) technology companies to submit high-performance models to the government for security review before release.
The New York Times reported Tuesday, citing a White House announcement, that "the executive order Trump signed is intended to solidify U.S. AI dominance and protect infrastructure and national security systems from cyber threats," adding that "the core content is establishing a government pre-release review framework for high-performance AI models developed by private companies."
The executive order is largely similar to the version Trump had planned to announce on November 21 but postponed, although the review period was shortened to 30 days from the original 90 days. Under the order, relevant agencies including the Treasury Department, the Department of Defense (War Department) and the Department of Homeland Security will establish a classified benchmarking process, while AI developers will form a voluntary cooperation framework with the government. Developers will determine whether models they are developing fall within the scope of review, and the government will be able to access the models for up to 30 days before release to check for security flaws.
With this measure, U.S. tech companies including Google, Microsoft (MS) and OpenAI welcomed the finalization of the amendment but are watching for future repercussions.
