Federal agencies directed to draw up AI safeguards for government use

The Biden Administration announced it is requiring federal agencies using artificial intelligence to adopt "concrete safeguards" by December 1, 2024, ensuring the protection of Americans’ rights and safety as the government expands AI use.

The Office of Management and Budget issued a directive to federal agencies to monitor, assess, and test AI’s impacts "on the public, mitigate the risks of algorithmic discrimination, and provide the public with transparency into how the government uses AI." Agencies must also conduct risk assessments and set operational and governance metrics.

The White House said agencies "will be required to implement concrete safeguards when using AI in a way that could impact Americans' rights or safety" including detailed public disclosures, so the public knows how and when AI is being used by the government.

New safeguards will ensure air travelers can opt out from Transportation Security Administration facial recognition use without delay in screening. When AI is used in federal healthcare to support diagnostics decisions a human must oversee "the process to verify the tools’ results."

The White House is requiring government agencies to release inventories of AI use cases, report metrics about AI use, and release government-owned AI code, models, and data if it does not pose risks.

Ongoing federal AI uses include the Federal Emergency Management Agency employing AI to assess structural hurricane damage, and the Centers for Disease Control and Prevention using AI to predict the spread of disease and detect opioid use. The Federal Aviation Administration is using AI to help "deconflict air traffic in major metropolitan areas to improve travel time."

The White House plans to hire 100 AI professionals to promote the safe use of AI and is requiring federal agencies to designate chief AI officers within 60 days.

In January, the Biden administration proposed requiring U.S. cloud companies to determine whether foreign entities are accessing U.S. data centers to train AI models through "know your customer" rules.