Biden Administration Is Set to Release AI Systems Guidance for Federal Agencies
A Biden administration executive order on artificial intelligence on deck for release next week will result in governmentwide standards for agencies already using the emerging technology, a top White House official said.
Arati Prabhakar, director of the White House Office of Science and Technology Policy, said Thursday the administration is aiming to “figure out what are all the actions that we can take to meet this AI moment” under existing law with the long-awaited executive order. President Joe Biden is expected to express support in the forthcoming guidance for the use of AI as a national security tool, while beginning to establish a comprehensive government framework for the safe and secure development and deployment of AI technologies.
“We are going to make really significant progress with the work that we’re doing from the executive branch – no question about that,” Prabhakar said at a Washington Post event. “And no question that much more will be needed.”
The National Institute of Standards and Technology is expected to play a key role in executing the order, from developing rules for testing and assessing AI systems for government use, to red teaming the systems that are employed by federal agencies. Prabhakar, who led NIST under the Clinton administration, said agency officials are “starting now to build out their capacity” around AI while expanding research partnerships with the private sector and academia.
The executive order will kick off a busy week for AI regulation around the world. Vice President Kamala Harris is scheduled to deliver a speech on Wednesday about the administration’s actions on AI development and regulation while in the United Kingdom for an AI summit (see: UK’s AI Safety Summit to Focus on Risk and Governance).
China, which has already introduced its own series of regulations on algorithms and AI, has an invitation to attend the summit. China’s inclusion has drawn backlash, including from former Conservative Prime Minister Liz Truss who Thursday demanded that Prime Minister Rishi Sunak withdraw the invitation. Truss’s tenure at 10 Downing Street is famous for losing an online competition over who would last longer – Truss, or a head of iceberg lettuce in a wig.
“The regime in Beijing has a fundamentally different attitude to the West about AI, seeing it as a means of state control and a tool for national security,” Truss wrote.
Prabhakar defended Chinese inclusion in the summit, telling Washington Post reporters at the event that “we always want to make sure that there’s good dialogue going on with every part of the world, so I thought that was a terrific idea.”
Senate Majority Leader Chuck Schumer, D-NY, told event attendees the Senate will host additional forums dedicated to exploring the impact AI will have on the workforce and election reforms.
It remains unclear what specific regulatory measures the administration may include in the executive order. The White House did not immediately respond to a request for comment.
Earlier this year, the administration received voluntary commitments from several leading AI developers to adopt safe, secure and transparent development processes for AI tools and technologies.
The commitments include critical cybersecurity steps to ensure the safe and ethical development of AI systems, including internal and external security testing, enhanced information sharing across the public and private sectors, and investing in insider threat safeguards to better protect sensitive data (see: IBM, Nvidia, Others Commit to Develop ‘Trustworthy’ AI).