The Biden administration is reportedly planning to apply new rules related to artificial intelligence on artificial intelligence and cloud companies that will result in their being required to report activities to the government.

The first new rule reportedly involves the White House planning to use the Defense Production Act, a U.S. federal law that allows the government to direct private industry to prioritize contracts and orders deemed necessary for national defense. It will require tech companies to let the government know when they train an AI model using a significant amount of computing power. According to a report from Wired, the rule could take effect as soon as this week.

The new requirement would give the government access to sensitive programs at companies such as OpenAI, Google LLC, Amazon Web Services Inc. and other companies with artificial intelligence programs. As part of the requirement, companies must also undertake safety testing on any new AI creations.

The proposed rule follows an executive order signed in October that established guidelines for the development of AI, including rules for the industry, security standards, consumer protections and federal oversight. The EO states that the government will approach privacy and safety in the development of AI from multiple angles, with the new rule part of the government’s approach to dealing with AI safety.

The use of the Defense Production Act to regulate AI safety, though, is an unusual move. Although AI has applications in military use, the tech sector is usually monitored through standard regulations, not an act created during the Korean War designed to expand the supply of materials and services from the U.S. industrial base to aid military and national defense.

The other new change on the horizon is a proposal that cloud companies must determine whether foreign entities are accessing U.S. data centers to train AI models, another side of AI safety. Reuters reported that Commerce Secretary Gina Raimondo said on Friday, “We can’t have non-state actors or China or folks who we don’t want accessing our cloud to train their models.”

To prevent China and others from accessing U.S. tech for AI, the administration published a proposal for “know your customer” regulation on Friday. The proposal would require cloud computing companies to verify the identity of foreigners who sign up for or maintain accounts that use U.S. cloud computing through a KYC program, similar to the requirements needed to open a bank account. The regulation also sets minimum standards for identifying foreign users and would require cloud computing firms to certify compliance annually.

Photo: White House