Face recognition is a complex technology that needs best practices and policies to ensure appropriate use, but those efforts are being hindered by inaccurate or biased statements. This paper provides experience-based insights to aid policymakers.
Face recognition is one of the most powerful and misunderstood technologies in modern times. In the hands of trained and experienced experts, it is properly and successfully used every day to make us safer. But it can also produce inconsistent results when operated with insufficiently trained individuals, or for improper purposes.
Best practices, policies, and oversight are needed to ensure the ultimate objective that we all desire: enabling appropriate applications with equitable outcomes while protecting civil rights and civil liberties, and an absence of nefariously or improperly-used applications. These ultimate outcomes are achievable, in the current day. Unfortunately, a significant literacy gap about the technology is keeping us from meeting this objective as many recommendations are inaccurate, biased, or based on hyperbolic assumptions. This gap must be overcome to produce legislation or policy that is evidence- and outcome-based, actionable, equitable, and measurable.