Columbia University researchers challenge the uniqueness of fingerprints, utilizing AI to analyze 60,000 prints. The AI claims 75-90% accuracy in linking prints from different fingers to one person, focusing on ridge orientation rather than traditional minutiae. Prof. Hod Lipson admits uncertainty about the AI’s workings, emphasizing its departure from traditional forensic markers. The study’s potential impact on biometrics and forensics is noted, though the tool’s current limitations for courtroom evidence are acknowledged. Despite skepticism, the study, set for publication in Science Advances, prompts consideration of fingerprint variability.