Louis v. SafeRent — $2.275 million settlement over algorithmic tenant scoring
What happened
Mary Louis and Monica Douglas, Black housing-voucher holders in Massachusetts, were denied apartments after SafeRent's 'SafeRent Score' rated them too low. The score leaned on credit history and non-housing debts and ignored that a voucher guarantees most of the rent. They sued in May 2022 (D. Mass., No. 1:22-cv-10800) on behalf of voucher holders and Black and Hispanic applicants. In July 2023 the court refused to dismiss, rejecting SafeRent's argument that a screening vendor is outside the Fair Housing Act.
The rule
Fair Housing Act and Massachusetts anti-discrimination law: disparate-impact liability for a scoring algorithm whose inputs disproportionately harm protected groups, applied to the vendor of the tool and not only to landlords.
Outcome
Settlement approved by the court on 20 November 2024: SafeRent pays $2.275 million (up to $1.175 million to class members) and, for five years, will not produce a SafeRent Score or accept/deny recommendation for applicants using housing vouchers unless a fair-housing expert validates a new model. The court awarded $1.1 million in attorneys' fees.
What this means for you
Scores built from credit data can replicate historic discrimination even with no protected attribute in the model. If you build or use scoring tools for housing, credit or insurance, test outcomes by race, ethnicity and income source, document the rationale for each input, and be able to switch off recommendations for groups the model treats unfairly.
Laws involved
- Colorado ADMT ActUpcomingUS-CO
- Quebec Law 25In forceCA-QC
- EU AI ActIn forceEU
Sources
- Louis v. SafeRent Solutions, LLC, 1:22-cv-10800 (D. Mass.) — Civil Rights Litigation Clearinghouse ↗
- Rental Applicants Win Ground-Breaking Settlement — Cohen Milstein ↗
- SafeRent settles case alleging algorithm discriminated by race and income — Fortune ↗
Last reviewed Sep 25, 2026.