In an AI-centric study led by a team at Columbia University, it was revealed that fingerprints might not be as unique as believed so far. Employing a deep contrastive network AI model to analyze over 60,000 fingerprints, the team established strong similarities among fingerprints from different fingers of the same person. This finding, though met with initial skepticism, has significant implications for forensic science and the law enforcement sector potentially improving investigations of cold cases and preventing unnecessary probes into innocent individuals.