Facial Recognition Software Increases Racial Profiling Against Black Individuals
Research shows that police facial recognition technology disproportionately misidentifies Black people, leading to wrongful arrests. The algorithms fail largely due to a lack of diverse training data and the amplification of existing human biases.
Police departments across the country use artificial intelligence-powered facial recognition technology to identify suspects, but this software routinely fails to accurately distinguish between Black individuals. Real-world consequences of these algorithmic flaws are severe, as evidenced by the wrongful arrests of Black men in Michigan and Georgia who face false theft charges simply because the system points officers to the wrong person.
Research reveals that law enforcement agencies utilizing this automated technology disproportionately arrest Black people compared to those that do not. This inequity stems directly from fundamental flaws in the technology, including a significant lack of Black faces in the algorithm training data sets and a dangerous, widespread belief among officers that the software provides infallible results.
Although automated face-matching offers potential benefits for public safety and efficiency, the severe risk of unconstitutional overreach demands immediate action. Without strict, enforceable safeguards in place, the integration of these flawed systems into policing simply magnifies existing human biases and guarantees further racial profiling.