By Rob Sherman, Deputy Chief Privacy Officer

The words “face recognition” can make some people feel uneasy, conjuring dystopian scenes from science fiction. Can someone use it to identify strangers on the street? Are institutions gathering mass databases of images that can be used to invade someone’s privacy or rights?

As government and non-government agencies, companies and others use face recognition technology in new ways, people want to understand how their privacy is being protected and what choices they have over how this technology is used.

Like many tools, face recognition can be used for good purposes — like helping people securely unlock their mobile devices, log into their bank accounts and make digital payments. It can help people organize their photos and share them with friends. It’s even being used to find missing and kidnapped children and to help officials confirm whether travelers have authentic passports.

But it can also be used in concerning ways. Some have raised concerns about how law enforcement uses the technology. Others have called attention to the potential for racial bias, arguing that facial recognition systems are more likely either to misidentify or fail to identify African Americans than people of other races. And while there have been proposals to regulate face recognition, there’s no consensus on whether or how to do so, and some approaches have been criticized for failing to focus on the most harmful potential uses.

This tension isn’t new. Society often welcomes the benefit of a new innovation while struggling to harness its potential. “Beware the Kodak,” one newspaper intoned in 1888 as inexpensive equipment came onto the market making photography available to the masses. They called it a “new terror for the picnic.” Confronting amateur photography for the first time, society could have restricted this technology – and fundamentally changed the way history was documented for more than a century. Instead, regulators took action on uses that raised concerns — for example, by prohibiting stalking or letting people sue for invasion of privacy — rather than requiring licenses to use “camera technology” or written consent forms before a person could appear in a photo. As a result, people became familiar with these early cameras, social norms evolved, and the world decided that the benefits of personal photography far outweighed the risks.

 Face Recognition and Facebook

On Facebook, face recognition helps people tag photos with the names of their friends. When you have face recognition enabled, our technology analyzes the pixels in photos you’re already tagged in and generates a string of numbers we call a template. When photos and videos are uploaded to our systems, we compare those images to the template. Here’s a video explaining how it works:

When we first introduced this feature in 2010, there was no industry standard for how people should be able to control face recognition.We decided to notify people on Facebook and provide a way to disable it in their account settings at any time.

We recently announced new features that use face recognition technology. People can now find photos of themselves even when they aren’t tagged in them, making it possible for people to manage their privacy in new ways. They may also know when someone is using their image as a profile photo — which can help stop impersonation. In addition, those with vision impairments can now hear aloud who’s in the photos they come across on Facebook. Just as in 2010, we had to evaluate how we’d inform people and give them choice over these new uses of the technology.

Our Responsibility

When it comes to face recognition, control matters. We listen carefully to feedback from people who use Facebook, as well as from experts in the field. We believe we have a responsibility to build these features in ways that deliver on the technology’s promise, while avoiding harmful ways that some might use it.

Our team has been working for more than a year to collect and respond to feedback on how people want to see us use this technology — and how we can do it most responsibly. People asked us to explain how face recognition works more clearly, and to provide more prominent information about how we might use it on Facebook. To address this feedback, we’re informing people about updates to face recognition in News Feed – the doorstep of Facebook.

We also decided to update Facebook’s settings. Concerns about updated settings are as old as Facebook, so we didn’t take the decision lightly. But we learned in our research that people want a way to completely turn off face recognition technology rather than on a feature-by-feature basis. We knew that as we introduced more features using this technology, most people would find it easier to manage one master setting rather than navigate a long list of products deciding what they want and what they don’t. Our new setting is an on/off switch. Some may criticize this as an “all or nothing” approach, but we believe this will prevent people from having to make additional decisions among potentially confusing options.

Finally, we aren’t introducing, and have no plans to introduce, features that tell strangers who you are. This was a common concern we heard from people when we researched new features that rely on face recognition technology.

 Moving Forward

As people use features on Facebook that use face recognition technology, we’ll learn more about what our community thinks of them — good, bad or in between. And we’ll learn what they think of our updated controls. We’ll build on these lessons and keep people informed about the work we’re doing to innovate and responsibly use this technology on Facebook.

Of course, it’s too early to know if face recognition will follow the path of the personal camera. But we look forward to the public’s feedback and to working with other companies and organizations as we continue to listen and learn.

Read more about our Hard Questions series.