In Washington and around the country, facial recognition technology has been touted by law enforcement agencies as a way to quickly identify suspects of different crimes. It uses algorithms to match photos from surveillance video to people’s photos from driver’s license databases and mugshots and identify people who have committed crimes. However, the technology is flawed and has resulted in false arrests.
Why facial recognition technology is flawed
Facial recognition technology relies on algorithms to measure and compare the distance between different facial features and match them to known photographs of people. In order to function properly, the algorithms must be trained with sample photos from a large and diverse group of people. When the algorithms were trained, the samples used did not include a sufficient number of photos of people in minority groups. As a result, facial recognition technology is prone to returning inaccurate results for black and brown people.
Facial recognition leads to false arrests
While several major companies have banned the use of facial recognition technology, businesses that sell this type of software have focused on the law enforcement market. Law enforcement agencies have used this technology to try to identify suspects in crimes despite the flaws in the algorithms. This has resulted in a number of false arrests. In one case, a black man was identified and arrested for stealing watches in front of his children. He was not told why he was being arrested and spent time in jail before he was able to prove that he was somewhere else at the time the crime was committed. His criminal charge is only one example of many in which people have been falsely identified and arrested for crimes they did not commit based on the use of facial recognition technology.
Despite the problems with this technology, law enforcement agencies continue to rely on it to make arrests. Courts should require police officers to base their arrests on more than a photo match using facial recognition technology. Given the known issues with this technology, facial recognition technology results should not be considered enough evidence to support probable cause for arrests.