Forensic analysis has long relied on the uniqueness of fingerprints as a reliable method for personal identification. However, according to recent research involving artificial intelligence, this uniqueness is now being called into question.
Under the guidance of senior student Gabe Guo from the Columbia Engineering School, a team applied artificial intelligence methods to analyze data containing approximately 60,000 fingerprints from a publicly accessible database maintained by the U.S. government. The results of this study were remarkable: fingerprints from different fingers of the same individual, while not identical, share a significant number of common features. Thus, despite individual differences, they can be identified as belonging to the same person.
The accuracy of this comparison method is currently modest, standing at 77% for a single pair of fingerprints. However, when multiple pairs are used, the accuracy of the comparison increases. Despite this, experts assert that the obtained results cannot yet serve as a basis for the use of this technology in the courtroom.
An article highlighting the research findings was published in the journal Science Advances. However, the path to publication was far from smooth. The researchers faced rejection from a well-known forensic journal, which dismissed their results, citing the long-standing concept of the uniqueness of fingerprints. This rejection was influenced by the reviewer’s adherence to the principle of “this cannot be because this cannot be.” After several attempts, the team reaffirmed their conclusions by providing additional data, and eventually, the article was accepted by Science Advances.
Despite the revolutionary results, skeptics have emerged. Professor Christoph Champod, a forensic science expert from the University of Lausanne in Switzerland, claims that nominal similarities between fingerprints from different fingers of the same person have long been known. He believes that the researchers have not presented anything new and have simply inflated the hype around their results, capitalizing on the popularity of artificial intelligence.