Forensics: This title is misleading. New AI discovery claims fingerprints aren’t truly unique — And it could rewrite forensic science

AI captured similarities fingerprint similarities in individuals not related to traditional loops, whorls and endpoints.

Human investigators have trusted fingerprints for more than 100 years — but what if a core belief behind that system was wrong all along ?
— Read on www.leravi.org/new-ai-discovery-claims-fingerprints-arent-truly-unique-and-it-could-rewrite-forensic-science-16521/

Unknown's avatar

About csidds

Dr. Michael Bowers is a long time forensic consultant in the US and international court systems.
This entry was posted in AAFS, Crime, Forensic Science and tagged , . Bookmark the permalink.

1 Response to Forensics: This title is misleading. New AI discovery claims fingerprints aren’t truly unique — And it could rewrite forensic science

  1. bredemarket's avatar bredemarket says:

    As I pointed out nearly two years ago after hearing a Double Loop Podcast on the topic, you need to differentiate between the hype surrounding the study and the actual study itself. If you read the actual study (“Unveiling intra-person fingerprint similarity via deep contrastive learning”), it refers to “similarity” but does NOT refer to “uniqueness.” https://www.science.org/doi/10.1126/sciadv.adi0329

Leave a reply to bredemarket Cancel reply