AI-based Study Finds Fingerprints are Not All Unique
“Using artificial intelligence, researchers analyzed 60,000 pairs of fingerprints from a government database and found several matches.”
by T.K. Randall
The findings could one day help to solve cold criminal cases and to acquit innocent people.
For the longest time, police investigators have used fingerprints to identify suspects and to tie them to crime scenes, leading to countless arrests and convictions.
This is because fingerprints are thought to be essentially unique to each individual finger, with no two being alike.
For this reason, fingerprints are also used as a form of biometric data to unlock everything from mobile phone lock screens to bank vaults.
But what if it turned out that fingerprints aren’t actually unique at all ?
Now according to a new study headed up by Columbia University engineering student Gabe Guo, fingerprints are not inherently unique – especially those of different fingers on the same hand.
Using artificial intelligence, researchers analyzed 60,000 pairs of fingerprints from a government database and found several matches. Some of these were from two different fingers belonging to the same person, while others were from two different individuals entirely.
Interestingly, the AI was able to determine with 77% accuracy which pairs of prints belonged to the same person, even though the researchers weren’t even sure how it was accomplishing this.
The findings could have implications in the field of forensic analysis because it could mean that, for example, unidentified prints found at a crime scene could belong to a known witness or suspect.
This, in turn, could help to link a suspect to more than one crime scene or to acquit an innocent person.
“Our tool is not good enough for deciding evidence in court cases but it is good for generating leads in forensics investigations,” said Guo.
by T.K. Randall