Anthropic has published a support article outlining when Claude may ask users to verify their identity. The policy is selective - not every session triggers a check, but certain features or contexts will prompt users to confirm who they are before proceeding.
The move follows a pattern that's becoming standard across AI platforms. Verification gates are appearing for specific capabilities, particularly where age restrictions, safety considerations, or compliance requirements apply. For most people doing everyday tasks - drafting copy, writing code, research - this change will likely go completely unnoticed.
Where it matters: developers building products on top of Claude, or team managers rolling out Claude access to employees, should know this policy exists before users hit an unexpected verification prompt mid-task. A sudden identity check during a long workflow can confuse users who weren't told to expect it. Anthropic's support documentation is the reference for specifics on which scenarios trigger verification and what the process actually requires.