US citizens have the right to opt-out of airport face recognition programs. This might cause a slight delay but maybe not. Even a small error rate would cause thousands of people to be misidentified every day so you might actually be speeding things up by skipping the surveillance.
In general, people are terrible at assessing risk. We intuitively associate familiar with safe and unfamiliar with dangerous. When bad things happen, we want someone to do something, even if it’s not helpful. This generally isn’t a good thing.
There have been a bunch of articles published recently about the terrible accuracy of facial recognition technology used by the police. This sounds really bad, but to understand it, you have to understand accuracy in biometrics.
A recent blog post mentioned that the Android and Apple phone specifications call for at least a 1 in 50,000 probability of a false match. In other words, a randomly selected imposter will have a 1 in 50,000 chance of unlocking your phone with his fingerprint. Let’s visualize that.
When people, companies, or organizations make a choice about a particular identity mechanism, they are making an economic decision. Every mechanism has a cost - and these costs should be balanced. A perfect identity system will have so much friction that it will allow no customers and a frictionless identity system will allow plenty of customers, but unlimited fraud.
Our co-founder Dr. Alex Kilpatrick has a lot of experience working with biometrics. The various modalities of biometrics (face, iris and fingerprint are the most common) have different applications. This video is a broad overview of the three most popular biometric modalities and common misconceptions about them.
Face matching is the easiest biometric for humans to identify with because we have done it literally from birth. And for the most part we are much better at it than computers - we can recognize faces with very little information, especially if they are familiar.
The retina is not a widely used biometric, because the process is painful and difficult. Most of the time, if you hear someone talking about retina scanning, they are actually talking about iris recognition.
We have spent a lot of time teaching computers how to recognize people. There are always cases where the computer performs surprisingly poorly and surprisingly well. The surprise is because computer “see” in a fundamentally different way from people.
With cameras on every street corner and the growing use of biometric technology in the consumer space, it's good to know how to avoid being recorded. This Ignite! presentation was originally given in 2010 at the Web 2.0 Expo in New York City.