US government to require AI companies to disclose safety test results

The government will now mandate that developers of significant artificial intelligence systems must disclose their safety test results to the Biden administration as part of their implementation process.

The White House’s artificial intelligence council is scheduled to meet Monday to review progress made on an executive order President Joe Biden signed three months ago to manage the rapidly evolving technology.

Chief among the 90-day goals of the order was an authorization under the Defense Production Act that requires AI companies to share vital information with the Commerce Department, including security tests.

Ben Buchanan, the White House’s special adviser on artificial intelligence, said in an interview that the government wants to “know that AI systems are safe before they’re released to the public — the president has been very clear that companies have to meet that requirement.”

Software companies have committed to categories of security tests, but companies do not yet have to adhere to a common standard for testing. The government’s National Institute of Standards and Technology is developing a unified framework for assessing security as part of an order signed by Biden in October.

Artificial intelligence has emerged as a leading economic and national security concern for the federal government, given the investment and uncertainties surrounding the launch of new AI tools that generate text, images and sounds, such as ChatGPT. The Biden administration is also reviewing legislation from Congress and working with other countries and the European Union on technology management rules.

The Commerce Department has developed a draft rule for US 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 completed risk assessments of the use of artificial intelligence in critical national infrastructure, such as the power grid.

The government has also increased the hiring of artificial intelligence experts and data scientists in federal agencies.

“We know that AI has transformative effects and potential,” Buchanan said. “We’re not trying to overturn the apple cart there, but we’re trying to make sure that regulators are ready to manage this technology.”