Anthropic Announces Identity Verification for Claude Starting July, Accelerating AI Industry Real-Name System

On June 10, 2026, Anthropic sent privacy policy update emails to some Claude users, announcing that starting July 8, consumer accounts (Free, Pro, Max) may be required to verify age or identity. Verification methods include uploading government-issued photo IDs (passport, driver's license, ID card, etc.) and taking a real-time selfie via third-party service Persona for comparison. Anthropic states that data is not stored on its own servers, not used for model training, and only for security and compliance purposes.

Background: AI Giants Tighten Identity Management Simultaneously

• OpenAI Leads the Way: In April 2025, OpenAI enabled document verification for API developers, accepting only physical documents, with each document limited to verifying one organization within 90 days. In January 2026, OpenAI quietly launched an age prediction model that cross-references signals such as account registration duration, active hours, and usage patterns to determine user age. Users flagged as minors are automatically restricted and must complete identity verification (upload documents or take a real-time selfie) to regain access. On June 17, 2026, some users reported a face-scan popup on the ChatGPT consumer side, requiring document upload and a 3D liveness detection video.
• Anthropic Follows: On June 10, 2026, Anthropic notified users via email of privacy policy updates, explicitly introducing identity and age verification. The verification scope has expanded from previously targeting minors, violators, or region-restricted accounts to broader consumer scenarios. The official support page indicates that as Agent tasks become more complex, verification scenarios may continue to expand.

Key Details: Verification Process and Data Flow

• Verification Methods: Users must upload a government-issued photo ID (passport, driver's license, ID card, etc.) and take a real-time selfie using a camera-equipped device for comparison with the ID photo.
• Third-Party Service: Verification is handled by Persona. Anthropic emphasizes that user data is not stored on its servers and is not used for model training.
• Trigger Scenarios: Includes using specific features, platform integrity checks, and security/compliance purposes. Anthropic specifically notes that as AI agent capabilities increase (e.g., executing multi-step tasks, connecting third-party apps), verification scenarios may further expand.
• Scope: Limited to consumer accounts (Free, Pro, Max); business customers (Team, Enterprise) are unaffected.

Reactions and Industry Impact

• User Reactions: Social media saw extensive discussion. Some users expressed privacy concerns, saying "This is no different from opening a bank account," while others suggested "Better switch to local models." Some users noted that emails appeared to be sent in batches, possibly indicating a gray-scale test.
• Industry Trend: OpenAI and Anthropic have both introduced identity verification measures within six months, marking the AI industry's shift from "anonymous use" to a "real-name system." Key drivers include:
  • Abuse Prevention: API-side document verification aims to prevent bulk registration, model exploitation, or training other models.
  • Agent Accountability: As AI evolves from chat tools to agents performing real-world actions (e.g., booking tickets, modifying files), platforms need to verify the true identity of instruction issuers to address legal and security issues.
• Data Security: Anthropic clarifies that verification data is processed by a third party and not used for model training, but user data may flow across platforms when connecting third-party apps.

Outlook

Anthropic's identity verification will take effect on July 8, with specific trigger conditions not yet fully disclosed. The industry widely believes that as AI agent capabilities continue to strengthen, the real-name system will become an industry standard, and users should prepare for the normalization of "face-scan verification" and "account cleanup."