S 3308
Artificial Intelligence Civil Rights Act of 2025
TL;DR
Senator Ed Markey (D-MA) introduced a sweeping federal bill that would regulate how companies use AI in consequential decisions like hiring, lending, housing, healthcare, and education. It requires pre-deployment bias testing, independent audits, human review of AI decisions, and bans certain uses like emotion recognition in workplaces and schools.
How This Might Impact Your Business
Companies using AI to make or inform 'consequential decisions' (hiring, firing, pay, credit, insurance, housing, healthcare, education, government benefits) would need to complete pre-deployment and annual algorithmic impact assessments testing for discrimination across protected classes.
Independent third-party audits would be required before deploying covered AI systems, with audit results submitted to the FTC and made publicly available.
Emotion recognition and biometric categorization systems would be banned in workplaces, schools, and housing, affecting vendors like HireVue-style interview tools and workplace monitoring software.
Individuals must be notified when AI is used in decisions about them and can request human review, meaning HR, lending, and insurance workflows would need new disclosure and appeal processes.
Enforcement runs through the FTC, state attorneys general, and a private right of action, so companies face lawsuits from individuals in addition to regulatory fines.
No small business carve-out is standard in Markey's civil rights framing, so mid-sized employers using off-the-shelf AI hiring or scoring tools would likely be covered alongside large enterprises.
Vendors and deployers share liability, meaning both the AI developer and the company using it can be held accountable for discriminatory outcomes.
What Should You Do
Inventory every AI or algorithmic tool your company uses in hiring, promotions, pay, lending, insurance underwriting, tenant screening, or customer eligibility decisions, and identify the vendor for each.
Ask vendors now for their bias testing documentation, training data disclosures, and whether they can support algorithmic impact assessments; if they cannot, start evaluating replacements.
Immediately review any workplace or customer-facing tools that analyze facial expressions, voice tone, or emotion, as these face outright prohibition under the bill.
Have legal and HR draft template notice-and-appeal language so you are ready to disclose AI use and offer human review if this or similar state laws (Colorado AI Act, NYC 144) advance.
Track the bill through the Senate Commerce Committee and monitor whether it gets a hearing; even if it stalls, its provisions are shaping state-level copycats.
Who It Affects
Sponsors
Status Timeline
committee
Read twice and referred to the Committee on Commerce, Science, and Transportation.
December 2, 2025