Can AI Make Forensics Better by Spotting Similarities in Fingerprints? A Fresh 'Discovery' Stirs Up Discussions

 If you enjoy watching crime TV shows, reading mystery novels, or listening to true-crime podcasts, you've likely heard that fingerprints are like personal signatures – each one is unique. Even identical twins, who share almost the same genetics, don't have identical fingerprints. This uniqueness plays a crucial role in forensic investigations, as matching fingerprints can link cases or unlock phones.

A team from Columbia University in the US believes that artificial intelligence (AI) might enhance the accuracy of forensic analysis. They developed a new AI system aimed at identifying similarities between fingerprints from different fingers of the same person.

Hod Lipson, a professor of engineering at Columbia University, shared the story of how the idea originated. A student approached him a year ago, challenging the notion of each finger having unique fingerprints. Lipson explained, "He fed pairs of fingerprints into a big AI system, some from the same person and some from different people." After a bit of training, the AI learned that fingerprints from different fingers of the same person are surprisingly similar, just requiring a different perspective.

When the researchers tried to publish their findings in a leading forensics journal, they faced rejection. However, the research eventually found its way into Science Advances. The team discovered that the AI system focused on the curvature of the ridges as the key factor in identifying similarities between fingerprints.

Christophe Champod, a forensic science professor at the University of Lausanne, Switzerland, and an expert on fingerprints, mentioned that these similarities were already known. He expressed skepticism about the AI system's practical use in forensics, stating that it might not significantly improve efficiency, especially when dealing with partial, distorted, or complicated traces at crime scenes.

Gabe Guo, the undergraduate student behind the research, acknowledged potential issues with degraded data due to smudged fingerprints at crime scenes. However, Champod and Guo both questioned the AI system's real-world applicability in forensic settings.

While Champod finds the AI-powered technique interesting, he doesn't see it as particularly useful for forensics. On the other hand, Lipson views the system as a glimpse into how AI could contribute to future discoveries. He believes this research showcases that even a relatively simple AI, given a plain dataset that the research community has had for years, can offer insights that eluded human experts for decades.

#machineminds #tech

are fingerprints unique plastic fingerprints fingerprints not unique iodine crystals the sharing economy is exemplified by when was fingerprinting invented ai fingerprints forensic science is the application of science to all of the following are ways in which dna technology plays a role in medicine except identifying mutations. producing insulin and growth hormone. tracking how a condition is passed between generations. examining paternity. how many nucleotides comprise a single strand of dna what is the purpose of codis the process of using an electric current to create a dna profile is called why are alleles helpful to forensic science while investigating a break-in, the police discover that the intruder was cut on the broken window that he or she used to enter the house and left a sample of blood behind. soon after the crime, the police apprehend a suspect and obtain a dna sample. which process is most likely to be used to determine whether the blood on the glass belongs to the suspect? father of fingerprinting what is the relationship between a child's alleles and the alleles of that child's mother and father collaborative consumption is when people get together to in forensics what is ibis #machineminds #tech