Blocking your face with your hand or hair won’t keep you from being tagged in a Facebook post. The company’s artificial intelligence lab cooked up an experimental algorithm that can recognize faces even when they’re obscured.
“There are a lot of cues we use. People have characteristic aspects, even if you look at them from the back,” Yann LeCun, Facebook’s director of artificial intelligence told New Scientist. “For example, you can recognize Mark Zuckerberg very easily, because he always wears a gray T-shirt.”
Facebook introduced its AI research project earlier this month. The team analyzed nearly 40,000 public Flickr photos with people’s faces in full and partial view using its algorithm that accurately connects photos to their likenesses 83 percent of the time.
Facial or pattern recognition software is everywhere: The technology has become so popular that some churches are using it to keep track of attendance. It’s used to find lost pets and a go-to tool for law enforcement agencies. The U.S. Customs and Border Protection tested it on thousands of unsuspecting travelers in Dulles International Airport.
Google made pattern recognition a central part of its Photos app, and while Facebook already uses it to suggest friends to be tagged in photos (whether they want to be or not), and plans to extend the technology to its private-sharing photo app Moments.
The use of facial recognition technology comes with a host of privacy problems, including what private companies like Facebook do with the data they collect, whether the government collects data and profiles innocent citizens. Facebook previously deactivated its facial recognition program in Europe amid privacy concerns, and is currently fighting a lawsuit on the matter. There are no federal regulations for pattern recognition software, and the government’s attempt to craft some is quickly unraveling.
The National Telecommunication and Information Administration, a division of the Commerce Department, has been working toward a draft voluntary guidelines for facial recognition software use since 2014 as part of the White House’s push for a consumer privacy bill of rights. But privacy groups began pulling their support of the process over concerns that individuals won’t be able to consent before the technology is used on them.
In a letter to the NTIA obtained by the Washington Post, nine privacy groups including the Center for Democracy and Technology, Electronic Frontier Foundation, American Civil Liberties Union, and Consumer Watchdog wrote:
“At this point, we do not believe that the NTIA process is likely to yield a set of privacy rules that offer adequate protections for the use of facial recognition technology,” the letter stated. “We are convinced that in many contexts, facial recognition of consumers should only occur when an individual has affirmatively decided to allow it to occur. In recent NTIA meetings, however, industry stakeholders were unable to agree on any concrete scenario where companies should employ facial recognition only with a consumer’s permission.”