A college student published a study which he says disproves the long-held belief that all fingerprints are unique, leading to pushback among experts.
Surasak Suwanmake/Getty Images
After an innocuous conversation about fingerprints with a college professor ahead of his freshman year, a now-senior at Columbia University set about using artificial intelligence to disrupt a long-held belief about the individuality of a human being's fingerprints.
Gabe Guo, the engineering student at the helm of the project, spoke with CNN about the new study published in the journal Science Advances this week, which he argues has disproven the popular assumption that every fingerprint on a hand is inherently unique.
The study relied on AI's discovery of similarities between different fingers belonging to the same person to argue that each fingerprint is, in fact, not totally unique.
The project, which was also a collaboration between Hod Lipson at Columbia Engineering and Wenyao Xu at the University of Buffalo SUNY, used an artificial intelligence model called a deep contrastive network, which is similar to the technology used for facial recognition, according to CNN.
The team input a US government database of 60,000 fingerprints in pairs. Some of the pairs came from two different fingers belonging to the same person, while other pairs were two different people's fingers.
The AI system discovered that fingerprints from different fingers belonging to the same person were incredibly similar. As the AI system improved its accuracy as the study progressed, it was eventually able to tell with 77% accuracy when a pair of prints belonged to the same person.
Researchers honed in on the angles and curvatures at the center of the fingerprint to find these similarities, Guo told CNN.
Guo emphasized to CNN the AI findings of the research, which he said could have far-reaching impacts beyond the niche world of fingerprints.
"This isn't just about forensics, it's about AI. Humans have been looking at fingerprints since we existed, but nobody ever noticed this similarity until we had our AI analyze it," Guo told the outlet.
Lipson, one of the study's co-authors, told Columbia Engineering that the findings could one day help revive cold cases and acquit innocent people.
"The most immediate application is it can help generate new leads for cold cases, where the fingerprints left at the crime scene are from different fingers than those on file," Guo told CNN.
The AI system's accuracy isn't sufficient enough yet to move the needle in a court case, Columbia Engineering reported, but it could help narrow down leads in tough cases.
MirageC/Getty Images
The study, however, is not without its critics. Guo told CNN that the team initially struggled to get the project published, saying the forensics community was hesitant to cast doubt on the belief that no two fingerprints are alike.
Forensics experts not involved in the study have expressed further doubts about the paper's novelty in comments to media outlets this week.
Christophe Champod, a professor of forensic science at the School of Criminal Justice of the University of Lausanne in Switzerland, told CNN that the study hasn't disproven anything radical, adding that the researchers' argument citing a correlation between fingerprints belonging to the same person has been well-known since the start of fingerprinting.
Meanwhile, Dr. Sarah Fieldhouse, an associate professor of forensic science at Staffordshire University, told The BBC that the study is unlikely to have a radical effect on criminal casework in its current stage.
Guo responded to the criticism by saying the study goes further than any research in the field has previously, CNN reported. The team also open-sourced their AI code to make it available to the public and other experts.
"I think this study is just the first domino in a huge sequence of these things," Guo told CNN. "We're going to see people using AI to discover things that were literally hiding in plain sight, right in front of our eyes, like our fingers."
