Sen. Edward Markey, D-Mass., and Rep. Yvette Clarke, D-N.Y., have reintroduced the Artificial Intelligence Civil Rights Act, which seeks to stop companies from using biased algorithms that influence decisions affecting civil rights and access to essential opportunities. The reintroduction was announced Tuesday by the office of Rep. Pramila Jayapal, D-Wash., one of the House co-leaders for the proposed legislation.
As lawmakers push new AI guardrails, government and industry leaders are preparing for major shifts in how AI is developed and deployed. The 2026 Artificial Intelligence Summit on March 19 will explore these evolving policy landscapes and what they mean for federal missions and public services. Register now to join the relevant discussion on AI.
What Types of Algorithmic Decisions Would the AI Civil Rights Act Cover?
The legislation applies to companies developing or using algorithms in critical decision-making areas, including housing, employment, lending, healthcare and education. It would prohibit algorithmic discrimination based on protected characteristics and require independently audited impact assessments before and after algorithm deployments.
The bill also requires transparency about when an algorithm is involved in a decision.
Why Are Guardrails on Algorithms Needed?
Bill sponsors referenced cases highlighted by civil rights organizations in which algorithms reflect inequities in the data used to build them. They said automated decisions are already shaping major life outcomes, and oversight is needed to ensure those systems do not replicate discrimination.
How Are Civil Rights Advocates Responding?
Supporting groups said many algorithms incorporate data influenced by longstanding inequities, which could result in unfair outcomes when applied at scale. They pointed to concerns involving mortgage approvals, hiring decisions and other processes where discrimination has been documented.
According to the Lawyers’ Committee for Civil Rights Under Law, the bill establishes accountability and transparency measures to help ensure algorithms do not automate bias. Other national advocacy organizations, labor groups and civil rights associations also voiced support for the standards included in the proposal.
