Claude Wants Your Face and Your ID Starting July 8 — Read This First

Security researchers found a government dashboard sitting on a public web address. The code behind it revealed something most people never expected. The company that verifies your identity for an A.I. chatbot can run nearly three hundred separate checks on you — including screening your name against terrorism watchlists.

Starting July eighth, the company behind the

Starting July eighth, the company behind the chatbot Claude can ask you for three things. Your government I.D. A live selfie. And a scan of your facial geometry — the unique map of measurements that makes your face yours. If you've ever typed a question into an A.I. tool, this story is about you. Anthropic, the company that makes Claude, is the first major American A.I. lab to write this kind of biometric collection directly into its consumer terms. So how did a chatbot end up asking for your face?

Let me start with the why. On June twelfth, a U.S. export control directive ordered Anthropic to block foreign nationals from its most powerful models. The problem? Anthropic had no way to check anyone's nationality. So the company pulled those models worldwide — then built identity verification afterward as the fix. The face scan isn't really about security. It started as a way to answer one question — what country are you from?

This didn't happen overnight. Anthropic quietly switched on biometric checks back in April, for what it called a few use cases. The July update spreads that across every tier of users. Three months — from a quiet test to company-wide policy. That's how normal gets built. One small rollout at a time, until handing over your face feels like just another login.

Who actually holds that data

Now, who actually holds that data? Not Anthropic. A third-party vendor called Persona. According to TechTimes, Persona can keep that information for up to three years — on its own servers, not Anthropic's. And remember that exposed dashboard? The code showed Persona can run two hundred sixty-nine distinct verification checks. That includes filing suspicious activity records with federal financial regulators. So the company verifying you for a chatbot has tools built for banks and watchlists.

Does this data really matter that much? Lawmakers think so. Under Illinois law, collecting someone's facial geometry without consent can cost a company between one thousand and five thousand dollars — per person, per violation. Facebook settled a case under that same law for six hundred fifty million dollars. That's how much regulatory weight one face scan can carry.

And here's the part that sits with me. If your password leaks, you change it. If your face leaks, you can't. You only get one. A central database full of facial scans becomes a target you can never reset.