Peter A Clarke

Facial recognition technology has been touted both as part of a dystopic nightmare or a leap forward in policing and enhanced security.  Without privacy safeguards the former is a likelihood.  As to how it will be used by police or security will determine how effective it is.  The real question however is whether it actually works.  To date the evidence has been at best mixed.  It does not work well with people who are wearing headgear.  Without a sufficient view of a target’s face it can be ineffective and it needs to be sufficient focused to work.  The industry touts its benefits and the media love facial recognition stories but it is technology which has had a long gestation and there is no guarantee of a safe delivery.

In Google Glass gets face recognition the AFR reports on an app being develop which will have face recognition glasses for Google Glasses.  The article provides:

Updated: Now this is why I would fork over $1500 and risk looking like a tool to wear a pair of Google Glasses: face recognition.

Lambda Labs, a San Francisco-based augmented reality company, says it has developed facial recognition software for Google’s not-so-soon-to-be-released Glasses, that will tell you the name of who you’re looking at. The software will be made available to Google Glass app developers within the week.

The software spots faces in your field of vision, and then puts names to that face, ranked in order of probability, based on a pool of mug shots that you’ve already uploaded to some servers somewhere. Presumably you can turn it off, so it’s not trying to name every passerby as you walk along the street, and you only turn it on when you arrive at a party or a meeting where you’re worried that you’ll forget a name.

Just how the software would work with Google Glass remains to be seen, however. Google has just issued a statement saying it’s reluctant to itself add face recognition software to Glass “unless we have strong privacy protections in place”, so it might be that you’d have to fiddle with your Android phone before the Lambda Labs software would run.

And by fiddle, I mean “root” – the unfortunate technical term for when you crack the security on a phone, allowing you to run unauthorised apps. That’s not hard to do on an Android phone, the main platform for Glass, and so despite Google’s protestations, the facial recognition genie will almost certainly get out of the Glass bottle one way or the other.

In case you haven’t been following, Google’s controversial Project Glass is a set of specs with a heads-up display built into it, that can act as a tiny, second display for apps running on your smartphone or tablet. The specs have a camera built into them, leading to concerns that their use will result in a massive, widespread privacy problem when people start shooting videos and stills all over the place without anyone’s knowledge or permission.

But none of that matters if the glasses help with facial recognition. The panic I get into when confronted with someone whose name I have forgotten, no price is too high to pay to save me from that.

Besides, face recognition could have positive social benefits, in a Panopticon, police-state sort of a way. Police in Britain are already contemplating the use of technologies like Google Glass as an investigative tool, and adding facial recognition to that mix could have a devastating effect on crime. Every time you walk past a wanted fugitive, an alert could pop up on your display, telling you to tackle the crim, or run for your life, depending on other pre-programmed variables such as your weight and size (dis)advantage, what sort of weapons the crim is known to carry, how much Dutch courage you’ve just consumed and so on.

It is interesting to see the reticence of Google to the proposal.  Google is something of a frequent flyer when it comes to privacy breaches.  That of course has not stopped it from developing specs which has a camera built in.