Claude Introduces KYC for Some Users

What has changed in access to Claude

This is a genuine tightening of access, but not a blanket ban affecting every new account. As of September 2026, Anthropic’s help center says identity verification may apply to selected users and organizations before they can access certain advanced Claude capabilities. The wording matters: it may apply; it is not mandatory for everyone.

Verification may require a physical government-issued photo ID and a phone camera for a live selfie. Organizations face a broader set of requirements: legal name, address, registration number, and proof of identity for the person submitting the request. Anthropic also links identity verification to applications for specific real-time cyber defense tools in Claude Opus and Sonnet.

The claim that older accounts keep working as before while all new accounts must pass KYC is more categorical than the official description. Available materials confirm the verification mechanism and its use in some scenarios, but do not support treating the line between old and new accounts as an absolute rule. That distinction between a user observation and a documented policy is important.

Based on the change description, the main barrier sits on the consumer access side, including Pro and Max subscriptions. API developers and business customers are identified as groups that are largely unaffected by this consumer gateway. That does not rule out organizational checks in specific processes, but it separates two distinct access paths.

Developers now face a different architectural risk

The main implication is straightforward: a consumer Claude account is a less dependable foundation for internal tools and automation. When a workflow relies on a user logging in manually, an additional ID document and selfie add friction at registration and may increase churn. For an experiment, that is an inconvenience; for a repeatable process, it becomes an operational risk.

An API-first approach appears more resilient because the available information does not place API access behind the same consumer barrier. The difference becomes especially clear at scale: a fleet of ordinary accounts now carries not only subscription limits, but also the risk of individual identity verification. This design was already fragile; KYC has simply made that fragility visible.

I would first clarify the verification boundaries: which features trigger it, what happens when a selfie check fails, and how organizational accounts are handled. Until these details are defined more clearly, the central issue is not KYC itself but the unpredictability of when it may appear in the middle of a workflow.

We previously covered how OpenAI API security triggers can alert account owners and require stronger compliance controls. Claude’s KYC requirement reflects the same move toward tighter identity and account governance for AI access.