Anthropic requires passport and selfie for some Claude users

Anthropic updated its help page on April 14 to require some Claude users to submit a government-issued photo ID and a live selfie through verification provider Persona. The company says verification records go to Persona’s systems and will be excluded from Anthropic’s model training.

The help page lists acceptable documents as a physical passport, driver’s license or national identity card. Photocopies, mobile IDs and student IDs are not accepted. Anthropic says a live selfie may be required to complete the check.

Anthropic selected Persona Identities to handle the identity checks and described itself as the data controller setting the terms of verification. ID images and selfies are sent to Persona’s systems, may be used to verify identity and to improve fraud detection, and are encrypted in transit and at rest, according to the published policy. Anthropic wrote, “We only use your verification data to confirm who you are and not for any other purposes.” The company also said the data will not be shared for marketing.

The company did not specify which Claude features will require verification or what specific user behavior might trigger a check. The help page says verification may appear when accessing “certain capabilities,” during “routine platform integrity checks,” or as part of “safety and compliance measures,” but it provides no further detail on thresholds or timing.

Anthropic gained large numbers of users in February after a rival signed a contract to deploy its AI on classified government networks; Anthropic reported record daily signups and a 60% increase in free users since January. Some users who moved to Anthropic cited privacy and safety concerns as reasons for switching.

Reaction to the verification rollout included criticism that the requirement is a company policy rather than a government mandate. Commenters noted that major competitors do not require government ID for access. Privacy advocates flagged risks in outsourcing sensitive documents to third parties and pointed to past breaches that exposed government IDs stored by other services. Observers also said a live selfie matched to a physical document can complicate access for users who rely on intermediaries in regions where Anthropic does not offer formal support.

Anthropic previously deployed content classifiers in December intended to detect users who identify as minors; several adult users reported account suspensions and loss of project histories while appealing incorrect flags. The company also enforces regional service restrictions that can lead to account bans for users registering from unsupported areas.

Anthropic described the verification program as part of broader platform integrity and safety work and confirmed verified data will be handled under the company’s stated limits. The company did not provide additional details about which Claude capabilities will require ID checks or how often users should expect verification prompts.