Our Fingerprints May Not Be Unique, Claims AI

Columbia University researchers are questioning the widely held belief that every fingerprint is entirely unique. In a recent study, a team from the US university used artificial intelligence (AI) to analyze sixty thousand fingerprints, aiming to determine if the AI tool could distinguish prints from the same person.

Researchers said the technology demonstrated an accuracy range of seventy-five to ninety percent in identifying whether fingerprints from different fingers belonged to the same individual.

Despite this success rate, Professor Hod Lipson, a roboticist overseeing the study at Columbia University, openly acknowledges that the team remains uncertain about the exact mechanism by which AI achieves this feat.

AI didn’t use traditional markers used for decades by forensics

The researchers suspect that the AI tool took a distinctive approach in analyzing fingerprints compared to traditional methods. Instead of focusing on minutiae—such as the way ridges end and fork—it seemed to concentrate on the orientation of ridges in the center of a finger.

“It is clear that it isn’t using traditional markers that forensics have been using for decades,” said Lipson. “It seems like it is using something like the curvature and the angle of the swirls in the center.”

AI flips the script on fingerprint lore – maybe they're not so unique after all

Discovery could have implications for the field of forensics The belief that all fingerprints are unique is so well accepted that crime novels and TV shows riff on it. Recent AI research has challe pic.twitter.com/5gJNtobfla

— John Marcellus (@safespace4space) January 11, 2024

Both Lipson and Gabe Guo, an undergraduate student involved in the study, expressed surprise at the findings. Graham Williams, a professor of forensic science at Hull University, said that the idea of fingerprints being entirely unique has never been definitively established.

“We don’t actually know that fingerprints are unique,” he said. “All we can say is that as far as we are aware, no two people have yet to demonstrate the same fingerprints.”

New research could impact biometrics and forensic science

Columbia University’s study findings might have far-reaching implications for both biometrics, involving using a specific finger for device unlocking or identification, and forensic science.

Consider a scenario in which an unknown thumbprint appears at crime scene A, and a distinct index fingerprint is discovered at crime scene B. These prints cannot be forensically linked to the same person. However, the AI tool could potentially bridge this gap.

Despite these potential breakthroughs, the Columbia University team, lacking backgrounds in forensics, openly acknowledged the need for additional research. AI tools typically need extensive data for training, and further development of this technology would require a larger fingerprint data set.

It’s important to note that all the fingerprints used to train the model were complete and of high quality, while real-world scenarios often involve partial or poor prints. “Our tool is not good enough for deciding evidence in court cases but it is good for generating leads in forensics investigations,” claimed Guo.