For over 100 years, police have believed that every fingerprint is unique. They use fingerprints to find and confirm suspects in criminal cases.
But a new study shows this may not always be true.
A team of researchers used artificial intelligence (AI) to study fingerprints. They found that fingerprints from different fingers of the same person can sometimes look alike. This challenges the long-held belief that each finger makes a completely unique print.
How AI Changed the Game
Hod Lipson from Columbia Engineering and Wenyao Xu from the University at Buffalo worked together on this research. A student named Gabe Guo led the study. He used a U.S. government database with about 60,000 fingerprints.
He trained an AI program to compare pairs of fingerprints. Some pairs were from the same person, while others were from different people. The AI was able to correctly guess whether two prints came from the same person 77% of the time.
When the AI looked at more than one print from a person, its accuracy increased even more.
A Hard Time Getting Accepted
At first, many fingerprint experts didn’t believe the results. A top forensics journal rejected the study. But the team kept working and improving their research.
Finally, their study was published in a trusted journal called Science Advances.
AI Sees What Humans Miss
Old fingerprint methods focus on tiny details like lines that split or end (called “minutiae”). But the AI looked at different parts of the fingerprint—like curves and loops—and found patterns people had missed.
The team believes this new approach could help police solve crimes faster, especially when multiple crime scenes are involved.
What’s Next?
The researchers know they need more fingerprints from a wider group of people to make sure the AI works fairly for everyone. They’re not saying AI should replace human experts. Instead, they want it to be used as a helpful tool.
This study shows that AI can help discover new things—even things we thought were already settled. And it also shows that you don’t need to be an expert to make a big impact.
As Lipson said, “We are about to see a lot of new discoveries made by non-experts using AI. The expert world needs to get ready.”
Source: earth
