AI proves that human fingerprints are NOT unique, sparking legal conundrum

In the realm of crime-solving, from television dramas like “Law and Order” and “CSI” to real-world investigations, fingerprints have long been the cornerstone for linking suspects to crimes. Now, AI may put an end to fingerprint convictions.

Fingerprinting has its limitations. Particularly challenging is the scenario where a criminal leaves prints from different fingers at separate crime scenes, making it difficult to connect these scenes and often leading investigations to a dead end.

Challenging traditional forensic beliefs

Conventional wisdom in forensic science has always held that fingerprints from different fingers of the same individual are distinct and cannot be matched.

This belief, however, has been fundamentally challenged by a fascinating study led by an unexpected source — Gabe Guo, a senior undergraduate at Columbia Engineering with no prior experience in forensics.

Guo’s journey began with a public database from the U.S. government, containing around 60,000 fingerprints.

By employing an artificial intelligence-based system known as a deep contrastive network, he analyzed these prints in pairs — some from the same individuals (but different fingers), others from different individuals.

AI approach to fingerprint analysis

The AI system, enhanced by the team through modifications to a state-of-the-art framework, gradually improved its ability to discern whether seemingly unique fingerprints actually belonged to the same person.

Astonishingly, the accuracy of the system for single pairs of fingerprints reached 77%. This figure soared when multiple pairs were analyzed, suggesting a potential to revolutionize forensic efficiency by more than tenfold.

This breakthrough didn’t come without its initial setbacks. When the team first submitted their findings to a well-established forensics journal, they faced rejection.

The reviewers, anchored in the belief that every fingerprint is unique, dismissed the possibility of detecting similarities between fingerprints from the same individual.

Overcoming academic rejection and perseverance

Undeterred, the team, comprising members from both Columbia Engineering’s Creative Machines lab and the University at Buffalo, SUNY’s Embedded Sensors and Computing lab, persisted.

They refined their AI system with more data, and despite the forensics community’s skepticism, sought a broader audience for their manuscript.

Their perseverance paid off. After an appeal by Hod Lipson, the James and Sally Scapa Professor of Innovation and co-director of the Makerspace Facility at Columbia, emphasizing the importance of their findings, the paper was accepted by Science Advances.

Lipson highlighted the potential of this discovery in solving cold cases and exonerating innocent individuals, although he acknowledged that the system’s current accuracy level is not yet sufficient to decisively conclude a case.

“I don’t normally argue editorial decisions, but this finding was too important to ignore,” Lipson said. “If this information tips the balance, then I imagine that cold cases could be revived, and even that innocent people could be acquitted.”

Discovering new forensic markers

The most intriguing aspect of this research was uncovering what new forensic markers the AI used, ones that had eluded experts for decades.

As Guo revealed, the AI didn’t rely on traditional minutiae — the branchings and endpoints in fingerprint ridges — but rather on the angles and curvatures of the central swirls and loops in fingerprints.

“The AI was not using ‘minutiae,’ which are the branchings and endpoints in fingerprint ridges – the patterns used in traditional fingerprint comparison,” said Guo, who began the study as a first-year student at Columbia Engineering in 2021.

“Instead, it was using something else, related to the angles and curvatures of the swirls and loops in the center of the fingerprint.”

Implications of this AI fingerprint revelation

The implications of this study are vast, extending beyond forensics. Aniv Ray, a senior at Columbia Engineering, and Judah Goldfeder, a PhD student, anticipate even more significant results once the AI is trained on a larger dataset.

“Just imagine how well this will perform once it’s trained on millions, instead of thousands of fingerprints,” said Ray.

The team is also conscious of potential biases in their data and acknowledges the need for further validation across diverse demographics for practical application.

This discovery stands as a testament to the untapped potential of AI in scientific discovery. Lipson notes that AI, even in its simpler forms and with accessible datasets, can uncover insights that have eluded experts for decades.

Future of AI in non-expert scientific discovery

Lipson further underscores the democratization of scientific discovery, evidenced by an undergraduate student’s ability to challenge long-held beliefs using AI.

“Many people think that AI cannot really make new discoveries–that it just regurgitates knowledge,” Lipson said.

“But this research is an example of how even a fairly simple AI, given a fairly plain dataset that the research community has had lying around for years, can provide insights that have eluded experts for decades.” 

Lipson added, “Even more exciting is the fact that an undergraduate student, with no background in forensics whatsoever, can use AI to successfully challenge a widely held belief of an entire field. We are about to experience an explosion of AI-led scientific discovery by non-experts, and the expert community, including academia, needs to get ready.”  

In summary, as AI continues to evolve, it presents a new frontier in scientific inquiry, opening doors for non-experts to contribute to groundbreaking discoveries.

This case serves as a harbinger for an imminent surge in AI-driven scientific breakthroughs, urging the expert community and academia to prepare for a new era of innovation.

The full study was published in the journal Science Advances.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates.

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–