The Biden administration is initiating new regulations on companies developing and using artificial intelligence (AI) by invoking the Defense Production Act.

These regulations include mandatory government notification when training new AI models, and potential monitoring and restrictions by cloud companies on foreign entities using US data for AI training.

The focus on AI safety is to protect against model poisoning, theft by foreign entities, and to ensure compliance with AI safety rules by major tech companies like Google and Amazon.

The regulations stem from an executive order aimed at establishing standards for AI that cover safety, security, privacy, equity, civil rights, and promoting American leadership in AI innovation.

The White House AI Council is reviewing the progress of this executive order, which also requires AI companies to share important information with the Commerce Department.

The National Institute of Standards and Technology is tasked with developing a safety assessment framework, while the Commerce Department drafts rules for U.S. cloud services offered to foreign AI developers.

Federal agencies have conducted risk assessments for AI in critical national infrastructure, and the government is hiring more AI experts and data scientists to help regulate and manage AI's transformative potential responsibly.