Facial recognition is the biometric technique of identifying or verifying a person from their face image. Operationalized at scale through Clearview AI (a US company that scraped 30+ billion images from social media and built a face-search product sold to law enforcement), PimEyes (a consumer-facing face-search service), Amazon Rekognition, NEC, and dozens of national-government systems. The technology that makes a single photograph in any database trace back to a name through cross-referenced public images.
What it means in practice
The 2020 New York Times exposure of Clearview AI made the operational reality public: an officer with a phone takes a photo of a suspect, runs it through the Clearview app, gets back a list of probable matches with linked social-media URLs. The same capability exists in PimEyes for consumer access (around $30/month for unlimited searches). The asymmetry is structural: any photo of you that was ever public (LinkedIn, Facebook, conference speaker page, news article, dating-site profile) is now indexed and searchable by anyone with the price of a subscription. The defenses are limited at the technical level (face is not changeable like a password) and primarily operational (control which photos are public, control where you are photographed).
Who uses it, and against whom
Operated by: at least 3,100 US law-enforcement agencies as Clearview customers (per the company’s own marketing claims, partially verified by procurement records), the consumer market via PimEyes and similar, retail loss-prevention systems in stores, airport and venue security systems, social-media platforms for tagging suggestions, and national-government surveillance programs (China’s Skynet, Russia’s Sphere, Iran’s Smart City systems). Against whom: anyone who has ever had a photograph appear publicly. Investigative journalists protecting sources whose face was incidentally photographed, activists at protests where police capture overhead drone footage, domestic-violence survivors whose ex-partner uploads an old photo to PimEyes, all face the same retrieval surface.
What you can change today
Three actions. First, audit your public face footprint: search yourself on PimEyes (the basic search is free; results list the public images they have indexed), screenshot the results. Second, where the photos came from accounts you control, consider removing or restricting (LinkedIn profile photo, dating-app archive, old conference talks). Third, reduce future capture: do not stand front-facing for casual photos in public, ask not to be tagged in others’ photos, decline photo requests at events where the resulting image will be public. The defense is partial because removal does not reach all the third-party copies; what you can control is the rate of new additions.