Anthropic's new terms can ask for your ID or your facial geometry

What the policy actually authorizes

The headline reaction was to a line in Anthropic's privacy policy, published June 8, 2026 with a July 8 effective date. The relevant passage is specific about what verification can collect: "In certain circumstances, we may ask you to verify your age or identity," and depending on method that means "an image of your government-issued identity document and the information appearing on it… your image in photo or video form, facial geometry templates (which may be considered 'biometric data' in some jurisdictions); and the result of the verification."

The policy also says where data is de-identified, Anthropic "will not attempt to re-identify such information, except as permitted by law" — an "except as permitted by law" caveat a few readers in the thread parsed warily. What the policy does not do is name a verification vendor or explain why it's required — leaving the thread to fill the vacuum, much of it with passports, Peter Thiel jokes, and threats to move to local Qwen or GLM models.

The correction that reframed the outrage

Before the dystopia talk could run too far, a high-reputation commenter posted receipts that this is neither new nor hypothetical:

"This is an updated terms of service, but they've had ID verification for certain accounts and situations for months." — Aurornis, linking an April Wayback archive of the verification help page and two-month-old Reddit reports

That reframed the complaint from "a new gate" to "a quiet, consistent rollout" — which one reply argued was worse, not better. The other grounding voice laid out why a frontier lab might feel cornered into this:

"A disturbed person was basically attaboy'd by AI into a mass shooting… there is a move by multiple governments around the world to ban children from using AI. You can't meet that legal requirement without age verification." — a long-tenured commenter describing the regulatory double-bind

That is the genuine bind: governments are simultaneously demanding child-safety age-gates and post-incident accountability, and neither is achievable for anonymous accounts.

The export-control angle nobody expected

The most surprising branch tied identity verification to a different motive entirely — re-enabling Anthropic's restricted top model for verified users only. One commenter wondered if checks were a precursor to bringing Fable back for US nationals, and another spelled out how dangerous that gets at work:

"only defense contractors are accustomed to handling export controls… something as little as discussing the usage of Fable to a non-'US Person' in the cubicle next to yours could be a crime… Export controls are no joke." — colechristensen, on what export-controlled model access would mean in an office

The skeptical counter, also worth stating once: a paid service that already has your credit card collecting more identity data is not the same as a state-run social-credit system, however the comparison flatters the outrage. Where the thread lands — for a developer building on Claude — is a concrete fork: enterprise users now have to ask whether a verification pop-up demanding a face scan is coming to their seat, and whether an export-controlled tier would make the model a compliance hazard rather than a productivity one.