Facial recognition is getting more controversial
I always felt that our modern society was becoming more and more lackadaisical with our privacy and data. No-one really seemed to care about privacy or how their data was being used. Then came along the EU’s GDPR regulation and to some degree things changed.
I think the thing that changed was that we began to realise there was some value in our data, that that value could be greater to others than we thought, and actually, protecting that data wasn’t the primary concern of the people we were sharing it with.
But did it change our behaviour? Not much, most people still tick agree to Privacy Policies and Terms and Conditions, but at least the conversation is now in the open and people are thinking about it.
One of the biggest areas of debate is biometric data. In other words physical characteristics of our human being, such as our face, finger-print, voice, retina, or even our shoe size and gait.
But for this article, I wanted to look at facial recognition in particular as there was a recent scandal in London around Kings Cross that hit the headlines.
Facial recognition, but am I bothered?
When trying to understand behaviour and attitudes towards technology, I tend to look at my own behaviour first. As a lover of technology, surrounded by it everywhere, I think I’m quite ahead of the curve, quite open minded, but when it comes to facial recognition I’ve noticed I’m somewhat more cautious.
A couple of examples:
- When I went to the world’s largest convention on mobile technology, Mobile World Congress, earlier this year, I opted out of their fast-track entrance because it used facial recognition.
- A couple of months ago whilst visiting London, I needed to use the loo, so headed for the nearest bar I could spot in the Embankment area. On the door there was a sign warning that they use facial recognition on the premises. Despite my immediate and rather pressing need, I decided to move on and find another place.
So what am I bothered about? Well, it’s not because I’m up to no good, and it’s not really that I’m worried about my picture and name being found online (it’s a bit late for that, given my career), but I think what really spooks me is how easy it is for you to be tracked without knowing.
An experiment by the New York Times earlier this year really highlighted that. In just a few days and for a few hundred Dollars they were able to legally build a system that could identify and track people across New York City. If you haven’t read it, I really think you should hop over here to read how they went about it, and their shocking findings.
Long distance and little control
The other thing that differentiates facial recognition over other types of personally identifiable information is that, unlike fingerprints or voice, my face can be seen and recognised by cameras a long distance away – often without me even knowing about it.
The other thing is that in many countries, deliberately obscuring your face can, at best raise suspicion, and at worst contravene local laws.
When I started thinking about this a few years ago, I started to notice a rise in what’s called “anti-tech” or “anti-surveillance” clothing. An example of this can be seen in this image.
In this example, the idea is that when a facial recognition system spots you, it also spots lots and lots of patterns that also look like lots and lots of other faces (squint at the image and you’ll see it looks like a crowd of faces smiling at you .. spooky). This simple technique essentially overloads the system and it can’t reliably determine which is the real face, a real-life Cloak of Invisibility (for all you Harry Potter fans).
Since this first appeared on my radar around 2016/2017, I’ve also started to see you can buy these patterns on T-shirts now, too. Like this one … (disclaimer, it’s my birthday soon and I just asked for one as a gift from my family 🙂 )
Update: To find out how the T-shirt worked, watch my video here.
Can facial recognition technology be used for good?
Yes, there are many awesome examples, but it’s the regulation and security that needs to be addressed with urgency.
However, a direction I really like is that from Sensing Feeling. They use all the magic of computer vision to recognise faces and emotions without matching those faces to identities. So although their technology can determine where in the picture a face is, and look at that face for clues about the person’s feelings, it doesn’t do what we know as facial recognition.
Sensing Feeling‘s platform looks for much more than faces, too. By looking at the pose, gait, facial expression and much more, they are able to look for things like stress, fatigue, joy, confusion, engagement and attention. There are obvious commercial opportunities for this, such as in advertising and retail promotions, but there are some real do-good applications too. Such as suicide prevention on bridges or railways, fatigue in drivers or pilots, education improvement to understand if students are confused or enlightened.
Cool tech too.
From a technical point of view, this is really quite cool. The solution uses cutting edge technologies such as machine learning and computer vision, on relatively small devices, with a relatively low connectivity requirement. In techie language we’d say it’s a distributed, edge-computing solution, on LPWAN IOT networks, AI-at-the-edge. Bazinga.