Claude Will Now Ask For Your ID. Here's What That Actually Means.

Claude Will Now Ask For Your ID. Here's What That Actually Means.

June 23, 2026 · 5 min read

Starting July 8, 2026, Anthropic can require some Claude users to upload a government-issued photo ID and submit a live selfie before continuing to use the service. This is a recently published policy change in Claude's own Help Center.

If you haven't seen this yet, here's the plain version.

What's actually changing

Anthropic updated its privacy policy to introduce a new category of data it can collect: "Verification Data." According to Claude's official help article on identity verification, this can include:

  • A government-issued photo ID (passport, driver's license, or national ID card)
  • A live selfie taken with your phone or webcam
  • A facial geometry template generated from that selfie, used to match it against your ID

The verification itself doesn't happen inside Anthropic's own systems. It's handled by a third-party vendor, Persona Identities, who collects and holds the ID and selfie on their servers. Anthropic says it can access those records when needed, for example to review an appeal, but doesn't copy or store the images itself.

This applies to Free, Pro, and Max consumer accounts. Business customers on Team, Enterprise, and API plans are not affected. Every time a policy like this lands, people start asking what it would take to just run AI privately, on their own machine, instead of waiting to see what the next update requires of them.

Who this hits, and when

The policy takes effect July 8, 2026. Anthropic has said the requirement currently applies to "a few use cases" and "a small subset of users," typically triggered by things like accessing certain higher-risk capabilities, patterns that resemble fraud or automated abuse, or accounts flagged for a possible policy violation where verification offers a path to appeal instead of an outright ban.

What hasn't been published: a full list of triggers, a specific data retention period, or a guarantee that the scope won't expand. Anthropic has stated identity data isn't used to train models and isn't sold or shared for marketing. It has also said that as Claude takes on more complex, multi-step tasks on a person's behalf, the need to confirm who's actually behind the keyboard is likely to grow, not shrink.

Translation: this is something most people are not prepared for. It is exactly this kind of moment that pushes people to shift toward local solutions they actually control. That is why the manual was written.

7 things most people teaching locally owned, private AI aren't telling you.

The truth about hallucinations. What security actually means at the machine level. What “local” really is — and what it isn't. Written for people who want to see clearly before they buy anything.

The choice between cloud AI and private AI is not a technology question but an ownership question. The technology works either way. The difference is who controls it, who sees what passes through it, and who gets to decide the terms of your access tomorrow.

The Private AI Manual is the complete build sequence for the machine that answers only to you. First edition. Available now.