AI Challenges Belief in Unique Fingerprints, Columbia University Study Reveals
In a surprising turn, Columbia University researchers trained an AI tool to analyze 60,000 fingerprints, challenging the assumption of each person having a unique fingerprint.
The technology claims 75-90% accuracy in identifying whether prints from different fingers belong to the same person.
The AI seems to focus on ridge orientation rather than traditional minutiae, leaving researchers unsure of its working mechanism.
While not yet suitable for court evidence, the tool could aid forensic investigations. The study, soon to be published in Science Advances, prompts a rethink on biometrics and forensic science.
However, experts caution about uncertainties in the AI’s markers and their consistency over time.
Source: BBC
https://www.bbc.com/news/technology-67944537