Chapter 34

You can't say the lock itself isn't secure; you can only say there's something wrong with the door. So the logic still needs to be clearly distinguished. And if you're thinking of cracking it at the level of the facial-recognition algorithm itself, that would actually require an even more astronomical amount of computation. Like with passwords, brute-forcing it would probably require several quantum computers—and we may have to wait decades for that. The cost would be enormous. Back when accuracy was at 60%, cracking it was still somewhat possible, but now it is almost impossible. So during the process of being cracked, what we should really be concerned about is not the security of facial recognition itself, but the security of its application. And one more thing: precisely because biometric traits are unique, you can't exactly change your face, can you?

Its inherent security is higher, so when users use it for security encryption, they place even more trust in it to protect more valuable private information. But conversely, if it is cracked—if we are ever cracked—then we may be compromised for life, because your face will not change. If its other security measures are neglected and it is ultimately cracked, the consequences would be even more severe, and the loss would be irreparable. You can't replace your face. Once it has been used, there is no way to take it back. Is that why, with biometric WeChat authentication, we would all need plastic surgery to keep using the original encryption method?

Right? Only then could we continue using it as the original encryption method. In today's flourishing age of facial recognition, fingerprint recognition, iris recognition, and other biometric technologies, we should think twice before acting, because once something goes wrong, the severity of the consequences will absolutely be multiplied.