Identity Verification Requirements for Claude Users

Anthropic has introduced identity verification on Claude to prevent abuse and comply with legal standards. Users must submit a physical government ID and a live selfie through a secure third-party partner called Persona. The system is designed with high privacy standards, ensuring data is encrypted and never used for training AI models.

Key Points

• Identity verification is required for certain use cases to ensure platform integrity and legal compliance.
• The process requires a physical government-issued ID and a live selfie, processed by third-party partner Persona.
• User data is protected via encryption and is strictly prohibited from being used for AI model training or marketing.
• Common reasons for verification failure include poor lighting or invalid ID types, with a manual appeal process available.
• Verified accounts are still subject to bans for violations such as underage use or accessing from unsupported locations.

Sentiment

The overall sentiment is strongly skeptical of Anthropic's identity verification approach. The community largely disagrees with the policy as presented, framing it as a privacy-invasive and strategically risky escalation rather than a routine safety measure. There is a meaningful pro-verification minority focused on accountability and abuse prevention, but it is outweighed by distrust of Persona, concern about government pressure, and resistance to identity-gated AI access.

In Agreement

• Verification can help Anthropic enforce bans, deter abuse, and reduce repeat misuse of powerful AI systems.
• Some users already provide identity-linked signals through payment, billing, email, and sensitive prompt content, so formal ID verification may add limited incremental exposure for them.
• Access to frontier models may reasonably require stronger accountability when misuse could create serious security or legal consequences.
• Age, residency, or eligibility checks may be necessary if laws or export-control obligations require Anthropic to restrict access.
• A few commenters prefer verification if it enables continued access to more capable models rather than blanket restrictions or product shutdowns.

Opposed

• Uploading government ID and a live selfie to Persona creates unacceptable privacy, breach, retention, subpoena, and vendor-trust risks.
• The policy may normalize real-name access to AI tools and contribute to a broader identity-gated internet.
• Non-US users fear that identity verification will combine with US regulation and export controls to restrict access to advanced American models.
• Identity checks may not stop serious adversaries because verified accounts can be bought, rented, compromised, or coerced.
• Payment methods and existing account signals already provide accountability, making full document verification disproportionate.
• Privacy-preserving proofs or limited eligibility attestations would be preferable to collecting full legal identity documents.