From 8 July, some Claude users will be asked to prove who they are. Under an updated privacy policy, Anthropic says it may require consumer users to verify their age and identity — potentially with a government-issued photo ID and a selfie — when they use "certain capabilities" or when the company runs safety and compliance checks. For a company that has built its entire brand on caution, it's a revealing move: a step that makes the platform safer and, at the same time, asks users to hand over more of themselves.

What's actually changing

The change is a clause in Anthropic's revised privacy policy, taking effect 8 July 2026. It gives the company the right to run age and identity checks on users of its consumer Claude plans — Free, Pro and Max. It explicitly does not apply to Claude Team, Enterprise, the Developer Platform, or anything governed by a separate commercial agreement.

In practice, that means an ordinary Claude subscriber could, from that date, hit a verification prompt where previously there was none. Anthropic frames it as a right it's reserving rather than a blanket requirement everyone will face on every login.

What triggers a check

This is the part worth watching, because Anthropic has been deliberately vague. It says a verification prompt may be triggered by three things: the use of "certain capabilities," platform-integrity reviews, or "safety and compliance" reasons. It has not published a list of which features or behaviours cross the line.

Reporting on the policy suggests the kind of moment that could prompt a check is an attempt to have Claude carry out a complex, multi-step task — the more capable, agent-style use that Anthropic is increasingly known for. But without a specific list, users won't know in advance what will trip the wire. That ambiguity is itself part of the story: "we may ask you to prove your identity when you do certain things, and we won't tell you exactly which things" is a meaningful amount of discretion to hold over a paying user.

What data is involved — and who holds it

Where a check does fire, it can be substantial. The policy describes collecting an image of a government-issued identity document, the personal details printed on it, facial images in photo or video form, and — notably — facial-geometry templates, the kind of biometric data that maps the unique structure of a face.

Anthropic's mitigation is that it doesn't do this itself. It uses a specialist identity-verification provider, Persona, to run the process. By Anthropic's account, the ID and selfie images are collected and stored by Persona rather than living on Anthropic's own systems, with Anthropic as the data controller and Persona as the processor. Whether that arrangement reassures you depends largely on how much you trust a third-party identity firm with your face and your passport — the data still exists; it's just held one step removed.

Why a safety lab is doing this

There are two readable motivations, and they reinforce each other. The first is age assurance: governments across the UK, EU and US are tightening rules that require online services to keep minors away from certain content and capabilities, and identity checks are the blunt instrument most platforms are reaching for. The second is misuse prevention: if a small set of advanced capabilities carry real-world risk, knowing who is using them — and being able to bar bad actors — is a lever a safety-focused company would want.

Both are coherent. The discomfort is in the trade. Anthropic has spent years positioning itself as the lab that takes the downsides of AI seriously. Asking users to submit biometric identity to access a chatbot's more powerful features is, depending on your view, either that same seriousness applied consistently — or a sign of how quickly "safety" can become a reason to collect more personal data than people expected to give a writing assistant.

The honest read

None of this is unique to Anthropic; age and identity verification is spreading across the whole internet, pushed by regulation as much as by company choice. And reserving the right to check ID is not the same as demanding it from everyone tomorrow. But the direction is clear, and it's worth users knowing before 8 July: the more capable these assistants become, and the more scrutiny they attract, the more likely it is that using them at full strength will mean proving, with a document and a photo of your face, exactly who you are.

If you're on a consumer Claude plan, the practical takeaway is simple — this won't change casual use overnight, but don't be surprised if a future complex request comes with a request of its own.