General Intelligence
Rite Aid’s Secret Facial Recognition System Is the Tip of the Iceberg
A stunning report on the company’s secret facial recognition program shows that banning government use of the technology isn’t enough

Welcome to General Intelligence, OneZero’s weekly dive into the A.I. news and research that matters.
Rite Aid stores have been using facial recognition software on customers in 200 stores, according to a Reuters investigation published this week.
These cameras were rolled out over the course of eight years and reportedly placed in lower-income and predominantly Black and Latinx neighborhoods, according to Reuters. They have also frequently misidentified people of color: A man stopped while shopping at Rite Aid even filed a complaint about algorithmic racial profiling with the California Department of Public Affairs in 2016.
The Reuters story paints a dark picture of an intrusive and flawed technology being developed, tested, and implemented in relative secrecy, and then quietly foisted onto customers in hundreds of locations across the country.
And private facial recognition in the United States doesn’t end with Rite Aid. Walmart uses facial recognition in more than 1,000 stores, 7-Eleven Australia uses facial recognition in 700 stores, and KFC and McDonald’s have tested the technology in China and Japan. There are likely many more examples of facial recognition being used in businesses across the United States, including algorithms used to track people in real time.
For instance, while facial recognition giant NEC told OneZero last year that it doesn’t sell real-time facial recognition to the U.S. government, it does have private customers in the United States that use the tech. The company would not disclose who.
The United States is maddeningly behind the pace of regulating this technology’s proliferation. Activists have won hard-fought victories against the government’s use of facial recognition in cities like San Francisco, Oakland, and Boston, and New York might ban it in schools. But that does nothing to stop companies like Walmart and Rite Aid.
In every state except Illinois, which has a strong privacy law protecting against the collection of facial recognition data, footage of your face can be stored, used, and sold without your consent or knowledge.
Rite Aid ostensibly holds data on millions of people who have walked through its stores, and what the company does with that information is up to its imagination. Though there are no public plans to do so, Rite Aid could conceivably spin off a business of selling its shoplifting blacklist to other stores.
Portland, Oregon, is the only city that has proposed a ban of the private use of facial recognition, which would mean projects like Rite Aid’s would be illegal. It’s a move that tech giant Amazon is working hard to oppose.
Other tech giants also would like to see their own hand in crafting facial recognition legislation, and their suggestions for new laws leave large cutouts for continuation of facial recognition in private businesses. A Microsoft employee who is also state senator in Washington wrote the state’s proposed bill on regulating the technology, and as you might expect, it was not very harsh on his employer’s business prospects.
Banning government use of facial recognition prevents the most obvious sci-fi dystopia of towns and cities under constant state surveillance. But without also banning the private use of the technology, we’d just be trading government surveillance for corporate surveillance, with even less transparency and oversight.
And now let’s check in on some of the most interesting A.I. papers of the week.
Google’s new Big Bird
Researchers propose a new way for text-analyzing algorithms to be able to read and retain more information—useful for question-answering algorithms and maybe even genomic analysis.
3D models from videos
Motion capture is becoming easier and easier, as Adobe and Stanford research shows how an algorithm can create a realistic, animated 3D model from a simple video.
An algorithm to murder mosquitoes
By automating the analysis of aerial footage, researchers hope to kill mosquitoes that spread disease faster and more efficiently. You gotta love it.






