Vice President Kamala Harris described the far-reaching nature of the role, commenting that chief AI officers will “oversee all — I’m going to emphasize that — all” use of AI within each agency.
She added that these officers must have “experience, expertise, and authority” to ensure that AI is used responsibly across the government.
According to a fact sheet, officials from various agencies have been meeting since December to coordinate efforts to ensure compliance with the new rules.
Federal agencies must also establish AI governance boards to coordinate and govern AI use by May 27. Only four agencies have created such a board as of March 28.
Other requirements
The White House also outlined various other requirements. Government agencies that use AI must introduce safeguards by Dec. 1, 2024. These requirements will require monitoring of public impact and will reduce algorithmic discrimination. Specific promises include rights protections related to healthcare, airport travel, and fraud detection in government services.
Federal agencies must also increase transparency around their use of AI. This requirement will require each agency to release an expanded annual inventory of use cases, report metrics otherwise withheld from the public, and release government-owned AI code and data when safe to do so. Agencies must also inform the public of any AI exempted from compliance with OMB policy.
Finally, the White House will take steps to grow the AI workforce, including by hiring 100 AI professionals in government by the end of July. The latest government budget includes an additional $5 billion for expanding a government-wide AI trading program.
The new requirements mean that federal agencies have completed the 150-day action in the Biden administration’s AI executive order from October 2023. The broader order describes many other AI safety, security, and privacy requirements.
