A new congressional bill would give federal agencies explicit authority to regulate AI systems when they contribute to violations of existing federal laws, closing a significant gap in how the government enforces protections against algorithmic discrimination. The Sectoral AI Governance Act, introduced by Rep. Sara Jacobs (CA-51), targets a real problem: AI is already making life-altering decisions for millions of Americans, yet federal regulators often lack clear guidance on how to apply existing laws to algorithmic systems.

Why Is This Gap Such a Big Problem?

AI systems are increasingly used in high-stakes decisions that directly affect people's lives. Whether someone gets approved for a mortgage, hired for a job, approved for health insurance, or accepted as a rental applicant often depends partly on algorithmic decision-making. The problem is that while federal laws already prohibit discrimination in these areas, agencies haven't always had a clear mandate to write rules specifically addressing how AI can violate those laws.

Consider a concrete example: if a rental screening algorithm trained on biased historical data systematically downgraded applicants from majority-Black or Hispanic zip codes, even when their income and credit scores matched those of approved applicants, that would violate the Fair Housing Act. But without explicit authority, the Department of Housing and Urban Development might struggle to require the company to test its algorithm for discriminatory patterns before deployment.

"Federal laws shouldn't become optional just because technology is new. AI is already helping make life-altering decisions for millions of Americans, but too often, it's operating in a gray area," said Rep. Sara Jacobs.

Rep. Sara Jacobs, U.S. Representative, California's 51st District

What Would the Bill Actually Do?

The Sectoral AI Governance Act takes a practical approach: rather than creating an entirely new regulatory framework from scratch, it leverages existing federal authorities. The bill would require agencies to write clear guidance on how current laws apply to algorithmic systems, giving them a consistent framework for issuing rules whenever AI is likely to materially contribute to violations of federal law.

The legislation includes several key safeguards to ensure the process works smoothly across government:

• Public Input Requirements: Agencies must seek early public comment through an Advance Notice of Proposed Rulemaking before proposing a rule, ensuring stakeholders have a voice in the process.
• Interagency Coordination: Agencies must consult with the White House Office of Information and Regulatory Affairs to identify overlaps, inconsistencies, or conflicts with other agency rules.
• Technical Expertise: Rules require technical consultation with the Office of Science and Technology Policy and the National Institute of Standards and Technology to ensure guidance is grounded in how AI actually works.
• Regular Review: Agencies must periodically review rules and amend or repeal them if they're no longer appropriate or properly tailored to the technology.
• Biennial Reporting: Agencies must report publicly every two years on how they use, or decline to use, this authority.
• State Authority Preserved: The bill protects states' ability to regulate algorithmic systems, except where state rules directly conflict with federal rules issued under the Act.

What Do Experts Say About This Approach?

The bill has drawn support from policy experts and civil rights organizations who see it as a pragmatic middle ground. Rather than waiting for entirely new AI-specific legislation, the approach leverages tools and authorities agencies already possess.

"Frameworks for responsible AI deployment do not need to be entirely built from scratch. While the federal government explores different avenues for AI rule-making, this bill highlights how agencies can use existing authorities to help tackle AI governance today," stated Owen J. Daniels, Associate Director of Analysis and Andrew W. Marshall Fellow at the Center for Security and Emerging Technology.

Owen J. Daniels, Associate Director of Analysis and Andrew W. Marshall Fellow at CSET

Brad Carson, President of the Alliance for Responsible AI, emphasized that the bill strikes a balance between innovation and accountability. "The Sectoral AI Governance Act is a smart, practical step toward making sure federal agencies can apply the laws already on the books when algorithmic systems contribute to real-world harms," he noted.

Brad Carson, President of the Alliance for Responsible AI

The American Civil Liberties Union also voiced support, arguing that as AI becomes embedded in housing, employment, healthcare, finance, and public services, agencies need clear authority to protect people from discrimination. The bill would require that workers, parents, and patients receive notice when AI is used in decisions affecting them, adding a transparency layer to algorithmic decision-making.

How Does This Fit Into Broader AI Governance Efforts?

This legislative approach reflects a growing recognition that AI governance doesn't require completely new legal frameworks. Instead, existing consumer protection, civil rights, and financial regulation laws can be adapted to address algorithmic harms. The bill essentially gives agencies permission and resources to modernize how they enforce laws that already exist.

The timing matters. As AI systems become more integrated into everyday decisions, the gap between what the law prohibits and what regulators can actually enforce has become increasingly visible. This bill attempts to close that gap by clarifying agency authority and requiring coordination across government to prevent conflicting rules.

The legislation also includes a limitation: enforcement would be limited to civil and administrative action by agencies, leaving unchanged the enforcement mechanisms available for violations of the underlying federal laws themselves. This means companies could still face lawsuits from individuals harmed by discriminatory algorithms, in addition to regulatory action.