Study challenges long-held belief: Not every fingerprint is unique

AI unveils astonishing fndings, potentially revolutionising forensic science


In a groundbreaking discovery that could transform the field of forensics, engineers at Columbia University have shattered a long-held belief that every fingerprint is unique. Their research, led by undergraduate senior Gabe Guo, has revealed that fingerprints from different fingers of the same person are surprisingly similar, challenging the very foundations of fingerprint analysis.

For decades, fingerprints have been considered the gold standard in linking criminals to crimes, both in popular culture and real-life investigations. However, when perpetrators leave prints from different fingers at separate crime scenes, connecting these scenes becomes an arduous task, and crucial leads often go cold.

The prevailing assumption in the forensics community has been that fingerprints from different fingers of the same individual, known as "intra-person fingerprints," are distinct and therefore impossible to match. However, Guo's team harnessed the power of artificial intelligence (AI) to question this widely accepted belief.

Utilising a public U.S. government database containing approximately 60,000 fingerprints, the team fed pairs of prints into a deep contrastive network, an AI-based system they had modified. Some pairs belonged to the same person but different fingers, while others belonged to different individuals altogether.

Over time, the AI system continually improved its ability to discern whether seemingly unique fingerprints belonged to the same person or not. The accuracy for a single pair reached an impressive 77%, and when presented with multiple pairs, the system's accuracy soared even higher. These findings could potentially increase current forensic efficiency by more than tenfold, offering a promising avenue for enhanced accuracy in criminal investigations.

To the team's surprise, when they shared their results with a well-established forensics journal, their findings were swiftly rejected. The anonymous reviewer and editor maintained that "every fingerprint is unique," dismissing the possibility of detecting similarities even if the prints originated from the same individual.

Undeterred by this setback, Guo and his colleagues persisted, providing their AI system with even more data. Remarkably, the system continued to improve, defying the skepticism of the forensics community. Determined to have their groundbreaking findings acknowledged, the team submitted their manuscript to a broader audience.

Although the paper faced further rejection, Hod Lipson, the James and Sally Scapa Professor of Innovation at Columbia Engineering, refused to let the discovery go unnoticed. Lipson appealed the decision, emphasizing the significant implications of the study. "If this information tips the balance, then I imagine that cold cases could be revived, and even that innocent people could be acquitted," he declared. While the system's accuracy is not yet sufficient to definitively determine a case, it shows immense potential in prioritizing leads in ambiguous situations. After further deliberation, Science Advances, a prestigious scientific journal, recognised the importance of the study and accepted the team's manuscript for publication.

This pioneering research challenges long-standing assumptions about the uniqueness of fingerprints, opening up new possibilities for forensic science. 

Join our WhatsApp Channel to get the latest news, exclusives and videos on WhatsApp