Machine learning is fast proving adept at overwhelming advanced security systems, as researchers have developed algorithms that can generate fake biometric credentials designed to trick bioidentification scanners.

Such easily produced synthetic fingerprints, facial profiles, iris patterns, or other bioidentifiers could be used to surmount any number of biometric scanner types.

Algorithms could be tailored to either replicate an existing match or use advanced computing speed to create near-instant generation of millions of fake biometrics, from which one might prove a match.