I shared some fingerprint-related information on my LinkedIn feed and other places, and I thought I’d share it here.
Along with an update.
You’re doing it wrong
Forensic examiners, YOU’RE DOING IT WRONG based on this bold claim:
“Columbia engineers have built a new AI that shatters a long-held belief in forensics–that fingerprints from different fingers of the same person are unique. It turns out they are similar, only we’ve been comparing fingerprints the wrong way!” (From Newswise)
Couple that claim with the initial rejection of the paper by multiple forensic journals because “it is well known that every fingerprint is unique” (apparently the reviewer never read the NAS report), and you have the makings of a sexy story.
Or do you?
And what is the paper’s basis for the claim that fingerprints from the same person are NOT unique?
““The AI was not using ‘minutiae,’ which are the branchings and endpoints in fingerprint ridges – the patterns used in traditional fingerprint comparison,” said Guo, who began the study as a first-year student at Columbia Engineering in 2021. “Instead, it was using something else, related to the angles and curvatures of the swirls and loops in the center of the fingerprint.”” (From Newswise)
Perhaps there are similarities in the patterns of the fingers at the center of a print, but that doesn’t negate the uniqueness of the bifurcations and ridge ending locations throughout the print. Guo’s method uses less of the distal fingerprint than traditional minutiae analysis.
But maybe there are forensic applications for this alternate print comparison technique, at least as an investigative lead. (Let me repeat that again: “investigative lead.”) Courtroom use will be limited because there is no AI equivalent to explain to the court how the comparison was made, and if any other expert AI algorithm would yield the same results.
Thoughts?
https://www.newswise.com/articles/ai-discovers-that-not-every-fingerprint-is-unique
The update
As I said, I shared the piece above to several places, including one frequented by forensic experts. One commenter in a private area offered the following observation, in part:
What was the validation process? Did they have a qualified latent print examiner confirm their data?
From a private source.
Before you dismiss the comment as reflecting a stick-in-the-mud forensic old fogey who does not recognize the great wisdom of our AI overlords, remember (as I noted above) that forensic experts are required to testify in court about things like this. If artificial intelligence is claimed to identify relationships between fingers from the same person, you’d better make really sure that this is true before someone is put to death.
I hate to repeat the phrase used by scientific study authors in search of more funding, but…
…more research is needed.
Bredemarket 
2 Comments