Artificial intelligence has discovered similarities between one person's fingerprints, something previously thought impossible

Fingerprints from the same person can have significant similarities previously thought impossible, according to a new study by a team of scientists at Columbia University. Until now, it was a well-established belief in forensic science that fingerprints from different fingers are unique even in the same person.

Here's What We Know

Using machine learning, the researchers analysed around 60,000 fingerprints from a public database. As a result, their neural network learnt to identify similar fingerprints belonging to the same person with 77% accuracy. This will potentially increase the efficiency of crime detection tenfold.

According to the authors of the paper, published in the journal Science Advances, their findings challenge the traditional views of forensic scientists. After initial scepticism, the paper was still able to be published.

The results show that even in long-established fields, artificial intelligence can make unexpected discoveries. Scientists expect a new wave of scientific breakthroughs using AI, including those involving non-specialists.

Source: TechXplore