Who's verifying your Claude account
I subscribe to Claude Pro. Starting July 8, Anthropic can ask me to prove who I am.
The takes circulating about this are mostly right and missing the more interesting part. Yes, Anthropic updated its privacy policy to allow identity verification. Yes, it can request your passport, a live selfie, and your facial geometry. Yes, a third-party vendor handles it.
But it’s not a universal requirement. It’s not “all Claude users must now show ID.” The policy applies to a small subset of flagged accounts - people whose accounts have been restricted but not banned, who can use the verification process to appeal. Commercial accounts - Team, Enterprise, API - are fully exempt. The immediate drivers are export control compliance, age verification laws, and fraud prevention. Not some broad surveillance mandate.
That’s the part the takes get mostly right. Here’s the part they’ve mostly missed.
The company processing your ID, your selfie, and your facial geometry is called Persona Identities. Persona is a San Francisco-based identity verification platform. It is backed by Founders Fund.
Founders Fund is Peter Thiel’s venture firm.
Founders Fund was also the first institutional investor in Palantir - the surveillance contractor whose ideology I’ve spent the last few months writing about in some detail. And Founders Fund invested $1.25 billion in Anthropic earlier this year, in its first direct position in the company.
The same venture firm is the first institutional backer of Palantir, a major investor in Anthropic, and the backer of the company that processes your biometric data for Anthropic’s identity checks.
I’m not saying this is a conspiracy. Investment portfolios overlap, and the same names appear everywhere in late-stage tech. Peter Thiel is not secretly engineering a surveillance panopticon out of Claude account verifications. Probably.
What I’m saying is that when the accountability layer gets built - the infrastructure that determines who gets in and who doesn’t, that holds your biometric data and your government ID - it keeps getting built by the same network of people. And that network has a specific set of beliefs about power, surveillance, and who should be in control. I’ve written about those beliefs at length.
Persona isn’t just asking “are you who you say you are.” Its platform can run 269 distinct verification checks - including screening against terrorism and espionage watchlists, monitoring adverse media across 14 categories, and filing Suspicious Activity Reports directly. Those are capabilities. Not all of them will be used for Claude account verification. But they exist, and they’re available.
Anthropic says your verification data won’t train models, won’t feed marketing, and stays between you, Persona, and Anthropic except where legally required. I mostly believe them. That’s also not the question.
The question is: when we build the infrastructure that mediates access to powerful AI systems, who builds it? Who owns it? What else can it do? What do its builders believe about power and who should wield it?
I’ve been making the same argument since April. The platform encodes ideology. The infrastructure enforces it. The people who control the accountability layer control more than access - they control the frame.
This is the accountability layer for AI getting closer to home. You should know who built it.