United StatesRetail and e-commerce2012–2023Rite Aid
In-store facial recognition for theft prevention
The FTC announced a five-year facial-recognition ban for Rite Aid after thousands of false matches
Problem
Rite Aid wanted to reduce losses from shoplifting and organized theft. It sought to alert staff when someone it considered likely to steal entered a store.
Approach
The FTC said Rite Aid used facial recognition in hundreds of stores from 2012 to 2020, matching shoppers against tens of thousands of images from security cameras, staff phones and news reports, many of poor quality. Thousands of false matches led staff to follow, search and publicly accuse customers. False matches were more common in stores in areas where Black or Asian residents were the largest population group. In December 2023, the FTC announced a settlement banning facial-recognition surveillance for five years and requiring deletion of the images and models built from them.
Results
Takeaway
Staff acted on false matches from a database containing many poor-quality photos. The FTC also faulted inadequate testing, monitoring and staff training.
Categories
- Size
- Large enterprise
- Industry
- Retail and e-commerce
- Business function
- Finance, risk and legalProduction and quality
- Innovation type
- Process innovation
- AI technology
- Computer vision
- Adoption
- Vendor solution
- Evidence
- Regulator, audit or court
Sources
Source summaries
The FTC announced a settlement banning Rite Aid from facial-recognition surveillance for five years. It alleged that the chain used tens of thousands of often poor-quality images in hundreds of stores from 2012 to 2020, producing thousands of false matches. False matches were more common in areas where Black or Asian residents were the largest population group. The settlement required deletion of images and derived algorithms, consumer notification and independent assessments.
Full details are in the original sources.
