Security is considered to be a right which everyone deserves. With over half the world’s population using smartphones as their primary means of computing manufacturers have to be extra cautious when it comes to smartphone security. Biometric authentication has been around for quite a while now with Apple’s iPhone 5S to be one of the very first devices to implement the technology which is considered pretty ordinary today. Manufacturers, however, have innovated a lot ever since with Samsung’s Iris scanner and Apple’s Face ID considered to be role models for modern-day biometric security.
The smartphone market has never slowed down and innovation just keeps on coming from places one wouldn’t even expect. Recent smartphones tend to have an under-screen sensor giving them a futuristic look. Fingerprint scanners are still considered to be the most secure means of encrypting information but that may soon change. Research from the New York University and the University of Michigan proves that with an AI software which has the ability to by-pass fingerprint security by adapting and improvising on the existing fingerprints till it decrypts the code. The AI is called “DeepMasterPrints” and consists of two working units. The first unit has the job to produce a 2D image of a fingerprint while the second unit differs it from others which allows the software to know if the required print matches the produced print.
Most smartphones, on the other hand, tend to not capture the complete fingerprints which further improves the chance for the AI bypass it. The researchers tested it out on three different levels of security with results ranging from 1.6% success rate to a staggering 76% success rate. This means that if 100 people used the most secure biometric scanner then the data of at least 12 of them will be exposed.
Apple when first announcing the removal of the fingerprint scanner many were not sure if it was the right step to make but looking at the statistics here they may have a point. Apple claimed to have a more secure encryption level by offering the unspoofable Apple ID which even with 3D models cannot be easily fooled. Hence it is proven that fingerprint encryption isn’t necessarily as secure as people may that and it only took a while for a legitimate workaround to be discovered. The facial recognition AI is still pretty new to the market and we may soon know if it lives up to the level of security manufacturers advertise.