A team of researchers from New York University has successfully used a neural network to create artificial fingerprints that can then be used to unlock devices that use biometric authentication systems–such as the fingerprint readers found in most smartphones, reports the Guardian.
The artificially generated fingerprints are dubbed “DeepMasterPrints” by the researchers and were able to imitate more than one in five fingerprints used in a biometric ID system. That system should have had an error rate of just one in a thousand.
DeepMasterPrints takes advantage of two flaws in biometric fingerprint systems. The first is that most fingerprint readers don’t image the entire fingerprint when scanned. Instead, a typical reader just looks for a match on just one part of the print. The second is some fingerprint features are more common than others, meaning systems that only scan partial prints are more likely to be fooled by common fingerprint characteristics.
Oh, and the DeepMasterPrints can not only fool biometric authentication systems but the human eye as well. So, about time everyone upgrades to a phone with Face ID then, right?