Facial recognition is a powerful, AI-driven surveillance technology used to identify individuals, often for law enforcement purposes. When facial recognition systems make mistakes, innocent people can be deprived of their liberty. This can happen when images used are grainy, the subject is not facing the camera directly, accuracy tolerances are relaxed, or law enforcement officials rely on a purported facial recognition match without sufficient investigation. Studies have uncovered bias in facial recognition: it has been shown to misidentify black people and women more often than white people and men. Using facial recognition to continuously scan a crowd to compare faces in the crowd to a watch list is often particularly inaccurate
Facial recognition can also be abused to surveil and to intimidate protesters. Immigration and Customs Enforcement (ICE) used facial recognition to identify people who protested against ICE abuse, including American citizens.
CDT advocates that when facial recognition is used for law enforcement purposes, it should be used only to investigate serious crimes, that a warrant should be required, and that persons arrested based in part on facial recognition matches be notified of its use.
