Facial recognition: Amazon adds 'fear' to range of emotions Rekognition can detect
Amazon is a major federal contractor. AWS powers ICE raids and detention camps.
Amazon is a major federal contractor. AWS powers ICE raids and detention camps.
Ring is Amazon's surveillance doorbell division, and a big part of their sales strategy involves terrifying people about the possibility of crime, partnering with police to assist in terrorizing Ring owners, and to provide police with warrantless, permanent, shareable access to surveillance doorbell footage (something the company has repeatedly lied about). — Read the rest
Between its line of Ring-brand surveillance doorbells and its "Rekognition" facial recognition product (both of which are used in law-enforcement and immigration-enforcement contexts), Amazon is at the center of the controversy over facial recognition technology.
One of the most obvious facts I've learned in covering the unfolding scandal of the secret deals between Amazon's Ring surveillance doorbell group and hundreds of US police departments is that Amazon loooooves to play word-games.
Amazon's Ring doorbells are surveillance devices that conduct round-the-clock video surveillance of your neighborhood, automatically flagging "suspicious" faces and bombarding you and your neighbors with alerts using an app called "Neighbors"; it's a marriage of Amazon's Internet of Things platform with its "Rekognition" facial recognition tool, which it has marketed aggressively to cities, law enforcement, ICE, businesses and everyday customers as a security measure that can help ID bad guys, despite the absence of a database identifying which faces belong to good people and which faces belong to bad people.
"BIG DEAL," says the ACLU's Matt Cagle about this story. "Amazon shareholders will vote on whether the Board must reconsider company sales of face surveillance to governments. The SEC rejected Amazon's attempt to prevent this proposal from moving forward."
Sources "familiar with Ring's practices" have told The Intercept that the company — a division of Amazon that makes streaming cameras designed to be mounted inside and outside your home — stores the video feeds from its customers' homes in unencrypted format and allows staff around the world to have essentially unfettered access to these videos.
Despite an uprising of Amazon employees over the use of the company's AI facial recognition program ("Rekognition") in law enforcement, the company is actively courting US Immigration and Customs Enforcement in the hopes that it will use the wildly inaccurate technology.
Tech-sector workers have enormous market-power: companies find it easier to raise cash than to use it to hire qualified developers.
After learning that Amazon was pushing the use of Rekognition, its facial recognition tool, for use in policing (a global phenomenon that is gaining momentum despite the material unsuitability of these tools in policing contexts), the ACLU of Northern California had a brainwave: they asked Rekognition to evaluate the faces of the 115th Congress of the United States.
Amazon bills its Rekognition image classification system as a "deep learning-based image and video analysis" system; it markets the system to US police forces for use in analyzing security camera footage, including feeds from police officers' bodycams.