So, here is a shocking fact for all the tech geeks. Now, Artificial Intelligence has the capability to create fake fingerprints to fool biometric identity scanners.
A group of researchers based in the New York University have recently found a network that has the ability to create artificial fingerprints to use it for authentication purposes on different smartphones and tablets to break their security. The new Neural Network that was suspected to be capable of this was called DeepMasterPrints and was able to copy fingerprints with a very less error rate i.e, 1:1000. The ratio of the matched fingerprint was 77 percent.
The team of researchers found 23 percent of a fingerprint replicated to what the network has created and this is enough to bypass the fingerprint scanners as they look for partial matches currently.
It was possible because of the flaw that was found in fingerprint sensors as they compare a small portion of the fingerprint for authentication that is stored in the database. As we know that the full fingerprints are unique, the copied portion of the fingerprint can bypass the dataset stored on the devices.
Photo: Getty Images / Wipada Wipawin
A group of researchers based in the New York University have recently found a network that has the ability to create artificial fingerprints to use it for authentication purposes on different smartphones and tablets to break their security. The new Neural Network that was suspected to be capable of this was called DeepMasterPrints and was able to copy fingerprints with a very less error rate i.e, 1:1000. The ratio of the matched fingerprint was 77 percent.
The team of researchers found 23 percent of a fingerprint replicated to what the network has created and this is enough to bypass the fingerprint scanners as they look for partial matches currently.
It was possible because of the flaw that was found in fingerprint sensors as they compare a small portion of the fingerprint for authentication that is stored in the database. As we know that the full fingerprints are unique, the copied portion of the fingerprint can bypass the dataset stored on the devices.
Photo: Getty Images / Wipada Wipawin
