CLEAR: fix identity layer to make fast AI safe AI

CLEAR frames the AI safety debate as a false choice: slowing AI down is not the answer, and neither is letting agents run on weak credentials. The piece points to a specific failure path. A fraudster impersonates an executive on an account-recovery call, takes over her account and registers an agent, after which fraudulent transactions follow. The company's security controls technically worked, because the agent faithfully carried out the fraudster's instructions. CLEAR argues that if true identity had been confirmed, the damage could have been stopped.

The problem traces back to documents and credentials that were never built to prove who you are. Driver's licenses were created to show you can drive and Social Security numbers to track wages, yet they became stand-ins for identity. Online verification made this worse, and generative AI can now supply a matching face and voice so that someone can know what you know, look like you and sound like you, and still be someone else.

CLEAR's proposed direction is proof of identity that travels with the person, through an immutable link between the person and identity and continuous confirmation everywhere they go. It does not describe a specific product or timeline. The commentary is signed by CLEAR's founder, and the views are presented as the author's own.