Rite Aid ran facial recognition for 8 years, never checked accuracy — FTC banned it for 5

Rite Aid ran facial-recognition surveillance in hundreds of stores from 2012 to 2020, matching customers' faces against internal watchlists of people the chain suspected of shoplifting or other misconduct. The FTC's complaint found the company never tested the system for accuracy before deploying it, never asked its vendors how accurate it was, and never checked afterward whether it was generating false matches.

It was. The system produced thousands of false-positive alerts, and employees acted on them: following flagged customers, searching them, ordering them to leave, and calling police over matches that were wrong. In one case cited by the FTC, a Black woman was accused of matching the alert profile of "a white lady with blonde hair." About 60 percent of the stores running the technology were in plurality-Black or -Asian neighborhoods, and the false-positive rate ran higher there than elsewhere.

The FTC's order, issued 19 December 2023, bars Rite Aid from using facial recognition for surveillance purposes for five years and requires it to delete any images and data collected through the program. The order carries no separate monetary fine, but Rite Aid had already filed for Chapter 11 bankruptcy earlier that year under the weight of opioid litigation and other liabilities, and the facial-recognition ban became one more mark on the record as it restructured.