AI companies required to report safety tests to US government

The Biden administration is set to implement a new requirement for developers of major artificial intelligence systems to disclose their safety test results to the government.

The White House AI Council is scheduled to meet Monday to review progress made on the executive order signed by President Joe Biden three months ago to manage the evolving technology.

One of the 90-day goals from the order was a mandate under the Defense Production Act for AI companies to share important information, including safety tests, with the Commerce Department.

Ben Buchanan, the White House special adviser on AI, emphasized in an interview that the government wants to ensure AI systems are safe before they are released to the public, as the president has made it clear that companies need to meet that standard.

While software companies have committed to a set of categories for safety tests, there is currently no common standard for the tests. The National Institute of Standards and Technology will develop a uniform framework for assessing safety, in line with the order signed by Biden in October.

AI has become a key economic and national security consideration for the federal government, especially with the launch of new AI tools like ChatGPT, which can generate text, images, and sounds. The Biden administration is also working on congressional legislation and collaborating with other countries and the European Union to establish regulations for managing the technology.

The Commerce Department has created a draft rule on U.S. cloud companies that provide servers to foreign AI developers.

Nine federal agencies, including the departments of Defense, Transportation, Treasury, and Health and Human Services, have conducted risk assessments regarding AI’s use in critical national infrastructure such as the electric grid.

The government has also increased hiring of AI experts and data scientists at federal agencies.

“We understand the transformative effects and potential of AI,” Buchanan said. “We’re not trying to disrupt that, but we want to ensure that regulators are prepared to manage this technology.”