In a groundbreaking study conducted by Columbia University, the long-held belief in the uniqueness of fingerprints is being reevaluated thanks to the influence of artificial intelligence (AI). A team from the university trained an AI tool to scrutinize a vast dataset of 60,000 fingerprints, aiming to determine whether it could accurately identify patterns linking fingerprints from the same individual.
According to the researchers, the AI technology demonstrated an impressive accuracy range between 75 and 90 percent in identifying whether fingerprints from distinct fingers belonged to the same person. However, the researchers are still grappling with the mystery of how AI accomplishes this feat.
Professor Hood Lipson, a roboticist at Columbia University who oversaw the study, admitted, "We don't know for sure how artificial intelligence does this." The team believes that the AI tool may be analyzing fingerprints in a novel way, focusing on factors such as the direction of ridges in the center of the finger, rather than conventional markers used in forensic analysis for decades.
"It seems to use something like curvature, the angle of the vortices in the center," explained Professor Lipson, highlighting the departure from traditional forensic methods.
The unexpected results surprised the researchers, leading them to double-check their findings. The study challenges the long-standing assumption that fingerprints are inherently unique. Graham Williams, professor of forensic sciences at the University of Hull, emphasized that while uniqueness was widely accepted, it was not conclusively proven, stating, "We don't actually know that fingerprints are unique, and all we can say is that as far as we know, no two people have yet shown the same fingerprints."
The implications of this study extend beyond theory, potentially impacting biometrics and forensic science. AI tools could become instrumental in connecting fingerprints from different crime scenes to the same individual, revolutionizing forensic investigations.
However, the Columbia University team, lacking a background in forensics, acknowledged the need for further research. AI tools typically require extensive data for training, and the study's reliance on complete, high-quality fingerprints may not fully reflect real-world scenarios where partial or inaccurate prints are common.
While the AI tool may not be deemed courtroom-ready, the researchers believe it holds promise in generating leads for forensic investigations. Sarah Fieldhouse, associate professor of forensic science at Staffordshire University, remains cautious about the study's immediate impact on criminal cases. She raised concerns about the stability of the features AI focuses on, considering variations in skin contact with printing surfaces and changes over a lifetime.
The peer-reviewed Columbia University study is set to be published in the journal Science Advances, marking a significant step in reevaluating our understanding of fingerprint uniqueness in the age of artificial intelligence.