Patrick Lambert looks at the increased use of facial recognition software.
Let me give you a little anecdote of something that happened to me last month. I walked into a local bank branch near where I live in Montreal, a bank where I don't have any account, and never interacted in any way, in order to ask about some investment products. I get directed to one of the staff members and I talked to her for about 10 minutes, never saying my name or giving her any way to identify me. I ended up not being interested in what she was offering and walked away. But as I did, the last thing she said was "Au revoir, Mr. Lambert." That's nice but how did she know who I was, despite the fact that I never gave any hint of my identity?
I didn't think much of it at the time, but after getting back home I did some research and found articles from a few years back about some banks here in Canada contracting with a company to provide them with facial recognition cameras. Going over the incident in my head, this seems like the most likely explanation. Now one argument people could make is that banks already have cameras, lots of them, and so do stores and buildings all around our modern towns. But I would argue that facial recognition brings this to a whole new level. What we're basically seeing is the same thing that happened years ago with phone calls.
Back in the day, when you picked up a telephone and placed a call, it would go along an analog circuit and the wire would need to be physically plugged in. We knew it was possible for someone at the central office to listen in, but the chance was remote, and there was no way they could start listening to every phone call as they went on. This is the same way normal cameras function. The best refinement the industry has managed to do is divide several cameras into a single screen, and then place multiple screens in front of a single security guard. But facial recognition completely changes the rules of the game. Now, we have software programs able to recognize anyone who walks by instantly, and record that in a database.
Facial recognition systems can work in a number of ways. The traditional system uses a standard camera and software that can compare some features of your face to a database of photos. By looking at the size of your nose, your eye position, or the slope of your jaw, it may be able to figure out who you are. A newer and much more reliable way to do facial recognition is to use multiple cameras and do a 3D comparison. This way, many more features of your face can be compared with what's in the database. Finally, some systems can even identify skin patterns such as lines, moles, spots and other characteristics found on most faces.
But how widespread is it really? This technology is nothing new. Back in 2004 the US State Department began implementing a huge facial recognition system for visa applicants. The New York Times reported last month that the Department of Homeland Security is in the final stages of developing such a system as well. The FBI is also spending over $1 billion on its own facial recognition system. Recently, it was revealed that the Attorney General of Ohio already implemented such a system back in June, without telling anyone. The system relies on 10 years of information from driver's licenses, gun permit holders, known criminals, sex offenders and more, all stored in a central database. The technology is also widely used in advertising. This video is just one such system which can access 36 million users within 1 second.
So we know the technology exists, and has become very refined. We also know corporations and governments are hard at work to implement it for their own purposes. The problem is that as with most things in technology, the law is lagging behind. There is no legal framework as to how facial recognition can be done. We're all used to seeing cameras everywhere we go, but the question that needs to be answered sooner rather than later is whether we're alright with these cameras automatically identifying us, recording everywhere we go, and then acting on this information, providing anything from personalized ads on a nearby billboard to a complete tracking profile at your local government office.
Have you had any experience implementing or using this kind of software?