A new study conducted by scientists at Columbia University found similarities in the fingerprints of different individuals by analyzing tens of thousands of fingerprints using artificial intelligence.
This finding could have important implications for biometric security and forensic science.
By examining more than 60,000 fingerprints, the research aimed to determine whether artificial intelligence could accurately match prints to individuals.
The tool demonstrated the ability to accurately identify tracks belonging to the same person with an accuracy rate of between 75% and 90%. However, researchers acknowledge that there is uncertainty about the methods AI uses to achieve these results.
It is thought that the AI may be using a new analytical strategy that analyzes the patterns and curves at the center of fingerprints, rather than focusing on the tiny details typically used in forensic analysis.
This approach is different from traditional methods and makes experts in the field curious.
Professor Hod Lipson, a robotics expert at Columbia University, acknowledged that the AI’s technique differs from traditional forensic applications in that it could possibly use the curvature and angle of the fingerprints’ rotations.
At the same time, Professor Graham Williams, a UK-based academic at the University of Hull, said he was not surprised by these findings given the lack of conclusive evidence that fingerprints are unique to each individual.
This groundbreaking research is poised to spark further research and debate within the scientific community.
Despite the study’s striking findings, it is not expected to disrupt judicial practices immediately.
The peer-reviewed study will soon be published in the journal Science Advances and will contribute to the ongoing debate on biometric identification.
