Algorithmic Accountability Act of 2025

Key Points

• Large companies would be required to perform ongoing impact assessments of AI and automated systems they use to make major life decisions, like who gets a job, a loan, housing, or healthcare. The FTC would write the detailed rules within two years, with another two years before companies must comply.
• The bill targets companies with more than $50 million in annual revenue or those handling personal data on more than one million people. Smaller companies with over $5 million in revenue are also covered if they build AI systems for use by the larger companies.
• Companies must test their AI systems for bias based on race, sex, age, disability, religion, veteran status, and other characteristics. They must try to fix any harm their systems cause and document what they did or explain why they could not.
• The FTC would create a searchable public website where anyone can look up which companies use automated decision systems, what decisions those systems make, and how consumers can contest or appeal those decisions.
• A new Bureau of Technology would be created inside the FTC with at least 50 staff experts in fields like machine learning, civil rights, and software engineering. State attorneys general would also gain the power to sue companies that violate the law.
• The bill does not override any state, tribal, or local AI laws, meaning states are free to pass even stricter rules on top of this federal framework.

Votes

No votes have been recorded for this legislation yet.