Facial recognition systems are quickly becoming a technological reality. Even social media giant Facebook jumped on-board the facial recognition bandwagon recently and is testing software that has recognition abilities that are, in practice, essentially as good as the human brain. If that isn't terrifying enough, consider the name of Facebook’s facial recognition project: DeepFace.
DeepFace uses something called deep learning, which is a form of artificial intelligence that forces machines to think more like the human brain, which uses a neural network to identify objects and create associations. The program works from photos that have forward-looking faces, so it begins by changing the angle of an image of a face until it meets that requirement. Then it starts looking for identifying features, matching what it knows about other faces to that data. If there are enough matches, DeepFace determines that it’s the same face, and identifies it as such.
DeepFace does this very, very well. In tests, it correctly identified faces 97.25 percent of the time. By comparison, humans only score slightly higher, at around 97.53 percent. This demonstrates that Facebook’s AI is almost as smart as we are, at least when looking at faces. So in the future, we probably won’t need to tag photos of our friends: Facebook will do it for us. This technology goes way beyond Facebook, though: everything from identifying criminals on CCTV to ID’ing terrorists at the airport can potentially benefit from this technology.
There are, obviously, huge privacy concerns that come with such facial recognition technology, so much so that an ex-Google CEO called the technology too dangerous. It’s also only a matter of time that this technology ends up in the hands of advertisers, who will scan our faces as we enter retail establishments and tell us what we want to buy. We must ask ourselves the question: is that a world we want to live in? And if it's not, is there anything that we can even do about it?
Via Extreme Tech