Anthropic has begun rolling out identity and business verification for selected uses of Claude, adding a formal check that may require individuals to present a government-issued photo ID and take a live selfie. The company says prompts may appear when people request certain capabilities, during routine platform-integrity checks, or as part of other safety and compliance measures. It has not described the measure as a universal requirement for every Claude account.
The verification process is operated with Persona Identities, which Anthropic selected as its partner. Individual users are told to have the original physical ID and a camera-equipped phone available. The document must be current, legible, undamaged and contain a photograph. Anthropic says scans, screenshots, photocopies, mobile IDs and nongovernment cards such as student or employee badges are not accepted.
Organizations face a related business-verification flow. A representative may need to provide the entity’s legal name, address and registration number, as well as their own qualifying photo ID and a live image. These requirements make the process more consequential than a conventional email or phone confirmation, while limiting it to information Anthropic says is necessary to establish identity.
According to the company, Persona collects and holds the ID and selfie images rather than storing those images on Anthropic’s own systems. Anthropic remains the data controller and can review verification records through Persona’s platform when needed, including during an appeal. It says transmitted and stored data is encrypted, fraud-prevention partners are contractually restricted to verification and related support, and identity information is not used for model training, advertising or marketing. Disclosure outside Anthropic and its verification partners would occur only when legally required, the company says.
A failed check does not necessarily mean an account is ineligible. Anthropic lists blurry images, unreadable or expired documents, incorrect company numbers and technical problems among possible causes. Users receive multiple attempts and are advised to improve lighting or use another eligible document before contacting support.
The help page says original physical government IDs from most countries are acceptable, but it does not publish a country list or promise that every document type will work.
Verification also does not guarantee continued access. Anthropic says an account can still be banned for reasons including repeated breaches of its usage policy or terms. Users who believe a ban was mistaken can submit an appeal to the company’s safeguards team. The rollout therefore combines identity assurance with Anthropic’s existing account-enforcement process, while leaving the exact set of capabilities and checks that trigger a prompt dependent on the company’s platform controls.
