When facial recognition technology is “properly constrained, (it) need not threaten privacy,” write Goodwin Partner Omer Tene and Xallient CEO and co-founder Lars Oleson. The process of verifying identification should not be a “dichotomy” between less risky one-to-one and riskier one-to-many verification, they contend, but instead be a “spectrum” of different methods, such as “one-to-a-few” verification, which can identify an individual biometrically, but only among a small, targeted group. They also provide an overview of AI products that use one-to-a-few verification, privacy through technical constraints and context-dependent recognition.
Full story