Police Used Facial Recognition Software To ID Suspects in UK Riots

We recently talked about how facial recognition technology was taking a turn for the scary and was being used to identify people in real life. That terrifying tech has already been used by cops and others in the London riots.

SlashGear is reporting that facial recognition technology was deployed in at least three different implementations last week in London as riots broke out. Some of these techniques may have been fairly successful, and some were definitely useless.

At the forefront was a program being used by Scotland Yard, which they've been quietly developing for use during the 2012 London Olympics. To what extent facial recognition was being used and at what percentage of success, we don't know as officials were unsurprisingly tight-lipped. What we do know is that the numbers of arrests continued to balloon well after the riots have subsided with roughly 1,635 total arrests and 940 individuals being charged. Exactly how this software works, we may never know. Government agencies are no fun that way.

Meanwhile, an independent group decided to put to use the exact type of technology we recently discussed. They formed a small group, developed some facial recognition software (largely, in this case, based on free code from Face.com), and cross-referenced riot photos with peoples' Facebook and other social network photos. In this case, the results the software generated were terrible enough for them to abandon the project. While this particular software wasn't so hot, it still drives home the point of our previous story: anyone can build this type of program, which uses this publicly accessible data, to identify you.

The last version of facial recognition being deployed is the old-fashioned kind that uses humans to do the processing. Various UK police agencies have been uploading photos of rioters to Flicker accounts and asking the good citizens of the realm to call in and identify the no-goodniks, should they think they recognize any. It's hard to believe that randos on their home computers trying to identify hooded subjects in grainy security-cam photos would generate particularly accurate results, but I'd tend to trust those results marginally more than those generated from a high tech computer program. Maybe.

The point is that not only is this stuff out there, it's being deployed already, by government agencies and random internet peoples. Again, there are no easy answers with regards to how to protect yourself from abuse by this technology, especially since so much of the data is publicly available. The first precaution I'd take, though, would be to change your social network privacy settings so that no one (who isn't already your friend) can see your profile photo. However, it's entirely possible that someone out there has already cached all of the publicly available info on their computers. Basically, this is the excuse you've been needing for that nose job you've always wanted. [CBS News and Forbes via SlashGear]


You can keep up with Brent Rose, the author of this post, on Google+ or Twitter.