Electronic devices rely on user verification techniques using biometrics such as facial recognition, but experts warn. Fingerprints may not be as effective as many imagine, the probability of two being the same is too high.
For decades it has been believed that fingerprints are the identity sign of each person, it is a kind of unique DNI. A recent study by Columbia University and the University of Buffalo in the United States shows the opposite.
Two identical footprints in the world
This biometric characteristic is not as unrepeatable as previously thought. The study published in the journal Science Advaced demonstrates with an effectiveness of 99.99% that Fingerprints from different fingers of the same hand have an enormous similarity.
The United States has analyzed a public database with 60,000 fingerprints. The results have been alarming, the study has discovered that the orientation of the central area of the prints ends up determining the similarity between different fingers of the same hand.
This discovery could forever change forensic science and digital authentication on devices, account logins, and payment platforms. A person’s fingerprints have a unique and identifying component.
Artificial intelligence in fingerprint recognition
Artificial intelligence has become a fundamental tool in fingerprint recognition. AI models can read repeating patterns and associate them with a user, as well as find similar traits of the same or another person.
The study clarifies that fingerprints are unique except in the very rare cases in which the same pattern can be repeated. Artificial intelligence used in biometric recognition methods you might be unable to detect that 0.01% difference between multiple fingers.
Furthermore, the study demonstrates that these systems are capable of determine when two prints belong to the same individual with an accuracy of 77%. If fingerprints were unique and unrepeatable this would not be possible.
The researchers who participated in the study have demonstrated something that was beginning to generate debate. The fingerprints of the same person are similar, but few imagined that they were similar enough to outwit fingerprint recognition systems.
The consequences for the moment are few since it is the same person, he will not start releasing criminals from prisons. For example, a mobile phone locked using this system could be unlocked with a fingerprint from another finger not assigned initially, but it would be indecipherable for another user, even a twin.