New Federal Requirement for AI Developers to Disclose Safety Tests

The Biden administration is taking significant steps to regulate the development of artificial intelligence (AI) by implementing a new requirement for major AI developers. This directive, stemming from an executive order signed by President Joe Biden three months ago, emphasizes the need for AI companies to disclose safety test results to the government. The White House AI Council is set to review the progress of this order.

A critical component of the order is a mandate under the Defense Production Act, obligating AI companies to share crucial information, including safety tests, with the Commerce Department. Ben Buchanan, the White House special adviser on AI, highlights the administration’s commitment to ensuring AI systems are safe before public release. Although there are designated categories for safety tests, a uniform standard is still in development by the National Institute of Standards and Technology.

AI’s emergence as a key economic and national security factor is evident, with the federal government focusing on managing new AI tools like ChatGPT. The administration is also considering congressional legislation and collaborating with international bodies, including the European Union, to regulate AI technology.

Furthermore, the Commerce Department is drafting rules for U.S. cloud companies providing services to foreign AI developers. Several federal agencies, including the departments of Defense, Transportation, Treasury, and Health and Human Services, have completed AI risk assessments for critical national infrastructure like the electric grid. This initiative is complemented by increased hiring of AI experts and data scientists in federal agencies, underscoring the government’s resolve to effectively manage AI technology and its transformative potential.