YouTube Announces Easy Face-Blurring Tool to Protect Anonymity of Video Subjects

Google's video site responds to a report from human-rights group WITNESS that criticized tech companies for not doing enough to allow protesters to mask their identities.


In recent years, viral videos -- particularly those of police brutality -- have been hailed for their ability to inspire mass protests. But activists have grown concerned that the documentary power of the Internet is a double-edged sword. The results of all this footage may not be just inspiration and communication, but repression, as states turn to the videos to identify and track protesters. As a result, fear of appearing in videos or photos of a protest may keep people out of the streets.

As the pseudonymous Rafeeq, a video-journalist and resident of Homs, wrote for Al Jazeera:

Many of my friends were arrested for protesting. However they weren't arrested from the protest sites, but rather from the checkpoints spread across the city.

But how did Assad forces know they protested?

Government forces have special teams dedicated to monitoring protests that we film and upload to the internet.

One of my friends who was detained for a short period told me that, as he was undergoing torture in detention, he was asked by the investigator if he ever participated in rallies against the regime. When my friend denied protesting, the investigator showed him footage where his face clearly appeared in a protest.

This is when we started learning how to film rallies from angles that would clearly show the crackdown by Assad forces on protests but not the faces of those protesting.

Concerned by incidents like these, the human-rights group WITNESS released a report last year calling on the big tech companies (naming specifically Google, Facebook, Microsoft, Apple, Yahoo, Twitter, and Nokia) to "step up their efforts in changing privacy controls, allowing for anonymity and developing user and content policies that better serve human rights defenders." They specifically noted that "No video-sharing site or hardware manufacturer currently offers users the option to blur faces or protect identity."

Today YouTube (part of Google) has responded to that need and announced an easy, in-site face-blurring tool. They explain on the You Tube blog:

Blurring faces on YouTube is simple. Once you've chosen the video that you'd like to edit within our Video Enhancements tool, go to Additional Features and click the "Apply" button below Blur All Faces. Before you publish, you will see a preview of what your video will look like with faces blurred. When you save the changes to your video, a new copy is created with the blurred faces. You will then be given the option to delete the original video.

This is emerging technology, which means it sometimes has difficulty detecting faces depending on the angle, lighting, obstructions and video quality. It's possible that certain faces or frames will not be blurred. If you are not satisfied with the accuracy of the blurring as you see it in the preview, you may wish to keep your video private.

A screenshot shows how the interface will appear:

Screen shot 2012-07-18 at 6.07.38 AM-png

YouTube is to be applauded for this effort and for working to give users a higher level of control over their content.

But the cynic in me can't help but see the limits of face-blurring. Yes it's great for protesters to be able to blur their own footage, but by no stretch of the imagination is that the only footage available to police. In Canada following the Stanley Cup Riots, citizens sent the Vancouver Police Department their videos and photos to help the police find the mayhem-makers. So many did so that it would take a 50-person team two years to analyze all the footage of an event that lasted just a few hours. Now, maybe riots over a sports loss are substantively different from protests in Homs, but how hard would it be for police to take their own video? Or for pro-regime sympathizers to do the same? Not to mention the hours and hours of footage constantly being recorded by security cameras on traffic lights and convenience stores.

In response to my question about this, Sam Gregory of WITNESS agreed in part, saying that, "This tool is not a panacea," but he reiterated that activists, at least will find the tool both useful and educational, and will help people not just in mass setting but in more direct testimonial videos in which they are the only actors. He said over email:

... Providing tools to citizens to control how citizen-shot footage can be shared more safely enables them to protect themselves better. In Iran in 2009, and now in Syria, people are being identified by security forces on the basis of footage shot by their fellow activists and protestors -- not only protests, but also when people talk to camera. Alongside this it's important to recognize that protests and public settings may in fact be the least prominent use of this type of tool -- as video becomes more common as a tool for speaking out, then it's often about a person giving testimony or speaking out in private (for example, a whistleblower or survivor of gender-based violence), and needing protection before that is released into the public sphere. ... We hope this is a great start to helping people take more control of their identities and protect themselves better, and that having tools like this from YouTube can better educate people who are filming their own content to think about the safety of themselves and the people in their footage.

In the end, the problems of surveillance, facial recognition technology, and state power run deeper than YouTube's face-blurring capacity can reach. But it's still an important step for Google to take, even if only for a limited set of scenarios. In the technological tug-of-war between government surveillance and activists' twin desires for both anonymity *and* a public platform, this tool is a valuable but small step in help activists attain those oft-competing aims.