Artificial intelligence now can make fake fingerprints which functions as a "primary key" for cellphones that have biometric sensors. The key made can be used for evil attackand security experts show that the performance percentage is quite low, but in some situations the system does work.
Namely, a measure of biometric security depending on the user's unique physical characteristics, like fingerprints, and eye shadows, but security researchers are increasingly proving it it is possible to deceive some, if not all, forms of biometric identification. In many cases, it is necessary to produce false biometric information so that the attack is successful, so that machine learning algorithms can also be made with fake fingerprints that can match a large number of fingerprints stored in the user's database.
This technique is known as DeepMasterPrintsand fake fingerprints They work like "main keys" that can open many doors. To make fake fingerprints, content creators are in a neural network upload the correct fingerprint from more than six thousand users. Synthetic fingerprints are then subject to tests, and if they are recognized by the system as a fake neural network, they will override minor corrections until the print is recognized as true.
Although synthetic molds still fail to deceive a large number of systems, they show the importance of once again examining security systems, and that Safety benefits of convenience identification method.