The National Security Agency is collecting millions of images intercepted from global communications for a facial-recognition program to identify and track intelligence targets, according to classified documents described by The New York Times.
The agency is using sophisticated software to harvest "millions of images per day" from emails, text messages, social media, videoconferences, and other communications, according to the documents. Once focused primarily on collecting telephone and email communications data, NSA officials believe the programs hold "tremendous untapped potential" that could revolutionize how the agency tracks surveillance targets, according to the documents, which were obtained by former NSA contractor Edward Snowden.
"It's not just the traditional communications we're after: It's taking a full-arsenal approach that digitally exploits the clues a target leaves behind in their regular activities on the net to compile biographic and biometric information" that can help "implement precision targeting," noted a 2010 document.
One NSA presentation described by the newspaper included several images of the same man in different settings and appearances, along with data points such as travel status and known associates. It wasn't clear how many images had been collected.
An NSA representative told CNET that the agency's foreign surveillance programs are designed to comply with US laws and policy direction.
"We would not be doing our job if we didn't seek ways to continuously improve the precision of signals intelligence activities -- aiming to counteract the efforts of valid foreign intelligence targets to disguise themselves or conceal plans to harm the United States and its allies," NSA spokesperson Vanee Vines said in a statement. "The lawful collection of foreign identity intelligence allows NSA to better identify and track such targets."
Although facial recognition technology has attracted growing attention in recent years from law enforcement and commercial interests, its reception has been rocky. Privacy advocates raised concerns in April over a facial-recognition database being developed by the FBI that could hold 52 million images by next year. While the FBI said the database could be a useful crime-fighting tool, the Electronic Frontier Foundation said its biggest concern with the database was the inclusion of face images for non-criminal purposes.
Facebook faced legal opposition from the German government over a controversial photo-tagging feature it rolled out in 2011. The social network is currently working on artificial intelligence software it says is capable of matching faces in images with nearly the same accuracy as humans. The DeepFace facial verification system uses a 3D modeling technique to detect faces, matching faces in large data sets with an accuracy rate of more than 97 percent.
The technology's application to commerce is also being explored. A Finnish company said last year it was developing a mobile payment system that would allow customers to complete transactions by having a point-of-sale camera snap a mug shot that could be compared to a database.