Cupertino, we have a problem. Face ID works. How is that a problem? Well, it’s not one for Apple because our favorite iPhone maker does facial recognition correctly. On device. Nowhere else. And with a touch of playful entertainment value. Typical Apple.
So, what’s the problem? Facial recognition is becoming a thing, a trend, and potentially a danger to anyone who prefers that nobody and nothing except their iPhone has an image of their face. Instagram and Facebook included.
Sidebar: Face ID works as well on my iPhone as Touch ID, but when it does not, it’s usually in the morning right after waking up. What does that say about quality of sleep? Maybe I need a new mattress.
It won’t be long and facial recognition devices will be everywhere, and somewhere along the line, some enterprising company will get in bed with politicians and we’ll have a national database of faces that continually gets updated thanks to traffic cams, Facebook, Instagram, and government efforts to peruse through photos and videos stored on iCloud and Google Drive.
No mocking, please. If you can think about it, it could happen. The technology is already available, but we don’t know if anyone is putting together all the pieces. Well, that’s not quite true. Some municipalities in China are doing exactly that. Everybody gets surveilled and the facial recognition is so good criminals can be picked out of crowds on the street.
How good and how pervasive will all that technology be in a decade?
If such capabilities bother you then get ready for clown makeup. Don’t laugh. Science often comes up with a way to counter new technology. Think cloud paint. It works, of course, because you’re painting your face in such a way that even sophisticated facial recognition won’t know who you are. Or, your dog. Or, spouse.
Some technology can modify a face on photos so that face recognition systems are fooled. It’s called adversarial training. Therein lies both the key to avoiding facial recognition technology and why it should be avoided. It’s a human adversary.
What about facial recognition that works in real time as you walk down the street, sit at a desk or coffee shop chair, or drive your car?
The aforementioned adversarial option may have legs. Technologists have created what are called adversarial glasses designed to fool facial recognition systems, or, in technospeak, facilitate misclassification.
I like that.
The eyeglasses were tested successfully against VGG and OpenFace deep neural network-based systems. Although the instructions for building them have not been made publicly available, the researchers say that the glasses could be 3D-printed by users.
I like that, too.
I understand the issues facial recognition brings to both sides of the party. Authorities love it because it can be instant and accurate. Humans doing nothing wrong probably don’t care. Some of us do. I consider it an invasion of privacy which a state sponsored camera can pick me out of a crowd, or follow me wherever I go.
Apple does Face ID right. Kudos and props. What about Google and Facebook? How are they protecting user faces where such captured information is part of their business model?