A judge has ordered SafeRent, an AI-powered tenant screening tool, to cease using its AI-generated scores to evaluate low-income renters, particularly those using housing vouchers. U.S. District Judge Angel Kelley approved a $2.3 million settlement stemming from allegations that SafeRent’s practices violated Massachusetts law and the Fair Housing Act.
How the SafeRent Scoring System Worked
The class-action lawsuit filed in 2022 claimed SafeRent’s scoring system unfairly harmed Black and Hispanic tenants and those relying on housing assistance. Plaintiffs alleged the system disproportionately assigned low scores to these groups, leading landlords to deny their applications. The settlement mandates significant changes to SafeRent’s scoring practices and compensates affected Massachusetts renters who faced housing rejections.
- SafeRent’s AI algorithm evaluated rental applicants using factors like credit history and non-rental debts, producing a “SafeRent Score” that landlords used to make leasing decisions.
- According to the lawsuit, this process lacked transparency, as landlords were not informed how scores were calculated.
- The system allegedly penalized applicants using housing vouchers, a practice criticized as discriminatory and in violation of housing laws.
Shennan Kavanagh, director of litigation at the National Consumer Law Center, explained that such scores draw on data designed to predict credit repayment rather than rent payments, making them unreliable for assessing tenant reliability.
Settlement Brings Major Reforms in AI Landlord Screening
As part of the five-year settlement, SafeRent will no longer use or display scores for applicants using housing vouchers nationwide. Additionally, landlords employing the “affordable” SafeRent Score model won’t see scores or recommendations to accept or deny tenants based on voucher use. Instead, they must evaluate applicants more holistically, considering their entire rental history.
SafeRent spokesperson Yazmin Lopez defended the company, stating that its scoring practices complied with the law. However, she acknowledged that litigation was costly and time-consuming.
“Defending the SRS Score would divert resources SafeRent could better use to serve its core mission,” Lopez said.
A Broader Trend of Scrutiny for AI in Housing
This case highlights the growing scrutiny of algorithmic decision-making in housing. In August, the Department of Justice sued RealPage, another property management software, for allegedly using AI pricing tools to inflate rents. The SafeRent settlement underscores the legal risks associated with using AI in housing decisions and calls into question the fairness of such systems.
The $2.3 million settlement represents a victory for fair housing advocates and sets a precedent for holding AI-driven tools accountable for discriminatory practices. The agreement aims to reduce bias in rental decisions and uphold equal access to housing opportunities by ensuring housing applicants are judged on their full records rather than opaque algorithms.