Eliminating Bias in Algorithmic Systems Act of 2026

Key Points

• Requires federal agencies that use, pay for, or oversee powerful computer decision tools to have a civil rights office focused on bias and discrimination.
• Defines covered tools broadly, including artificial intelligence, machine learning, and similar systems that can affect access to government programs, economic opportunities, or protected rights.
• Each covered agency must send Congress a public-style update every 2 years (starting within 1 year) on bias risks, what they’ve done to reduce harm, and what they recommend next.
• Agencies must talk with outside groups—industry, civil rights and consumer advocates, workers’ groups, technical experts, and affected communities—when looking at algorithm harms.
• Justice Department must set up a cross-agency working group so civil rights offices coordinate on algorithm bias and discrimination issues.