What laws currently govern how the US federal government can use AI

A Patchwork Rather Than a Single Comprehensive Law

There is currently no single, comprehensive federal statute specifically and exclusively governing all uses of AI by the U.S. federal government. Instead, federal government AI use is currently shaped by a patchwork of executive orders, agency-specific guidance, voluntary frameworks, and the continued application of existing laws to decisions that happen to involve AI.

The Role of Executive Orders and OMB Guidance

Presidential executive orders and guidance from the Office of Management and Budget have played a significant role in establishing requirements and expectations for how federal agencies should develop, evaluate, and deploy AI systems, including requirements around risk assessment, testing, and documentation for certain categories of government AI use — though this guidance can change as administrations and priorities shift over time.

NIST’s AI Risk Management Framework

The National Institute of Standards and Technology has developed an influential AI Risk Management Framework, providing structured guidance for identifying and managing risks associated with AI systems. While this framework is voluntary rather than a binding legal mandate, it has become a widely referenced standard that many federal agencies and other organizations use to guide their own AI risk management practices.

Why Existing Laws Still Apply to AI-Involved Decisions

Importantly, existing laws governing privacy, civil rights, and administrative due process continue to apply to government decisions even when AI systems are involved in making or informing those decisions — using AI doesn’t create a legal exemption from these existing protections, meaning a government decision that would violate civil rights or due process protections if made by a human remains legally problematic if made or informed by an AI system instead.

Why This Patchwork Approach Has Both Strengths and Limitations

This combination of executive guidance, voluntary frameworks, and existing law provides meaningful, if imperfect, structure for federal AI governance, offering flexibility to adapt as AI technology and understanding of its risks continues to evolve, but also creating some inconsistency and gaps compared to what a single, comprehensive federal AI statute might provide, which is a subject of ongoing policy debate and proposed legislation.

Why This Landscape Continues to Evolve

Given how rapidly AI technology and its government applications have been developing, this regulatory and policy landscape continues to evolve, with ongoing legislative proposals, updated executive guidance, and evolving agency practices all potentially reshaping the specific rules governing federal AI use over time.

Bottom Line

U.S. federal government AI use is currently governed by a combination of executive orders, agency-specific guidance from bodies like OMB, voluntary frameworks like NIST’s AI Risk Management Framework, and the continued application of existing privacy, civil rights, and due process laws — rather than a single, comprehensive federal statute specifically dedicated to regulating all government AI use.

Go deeper

• What regulations currently govern self-driving car testing and deployment?
• Is It Illegal to Create a Deepfake of Someone Without Consent?
• Do Regulations Require Disclosure of AI-Generated Advertising Content?