Track 1 platform and get the weekly governance digest. No credit card required.
This page describes what the document states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability may vary by jurisdiction. Methodology
This document establishes Anthropic's data collection and use practices for Claude.ai and related consumer products. Anthropic collects chat messages (user inputs and Claude outputs), device identifiers, IP addresses, browsing activity within the service, and payment information, with authorization to use this data for AI model training unless the user elects the opt-out setting in account preferences. The policy specifies that conversations flagged for safety review or submitted as explicit feedback may be retained and used for model training independent of opt-out status.
This Privacy Policy, effective January 12, 2026, governs Anthropic's collection, use, disclosure, and processing of personal data when Anthropic acts as a data controller across its consumer-facing Services (including Claude.ai) and Commercial Services (including Claude Team plan); it expressly excludes scenarios where Anthropic acts as a data processor on behalf of commercial customers. The policy states that Inputs (prompts), Outputs (AI responses), device identifiers, browsing activity, IP address-derived location data, and usage information are collected; the terms authorize use of Inputs and Outputs to train AI models unless users opt out, with a carve-out permitting training use even after opt-out when conversations are flagged for safety review or explicitly reported via feedback mechanisms. The opt-out carve-out for safety-flagged conversations is operationally distinct in that it permits continued model training use of user content without user consent in circumstances defined unilaterally by Anthropic, which may require evaluation under GDPR lawful basis requirements and CCPA-equivalent frameworks depending on jurisdiction. The policy engages GDPR (for EEA users), CCPA and California privacy statutes, Brazilian LGPD, Canadian PIPEDA or provincial equivalents, South Korean PIPA, and Washington State consumer health data law; regional supplemental disclosures are provided for Canada, Brazil, and the Republic of Korea, and a separate Consumer Health Data Privacy Policy applies to Washington State users who integrate third-party health applications.
Institutional analysis available with Compliance
Regulatory exposure by statute, material risk assessment, vendor due diligence action items, and enforcement precedent. Available on Compliance.
Start Compliance free trial1 important change detected
3 versions captured · Last updated: June 2026
New explicit disclosure of Anthropic's internet scraping and third-party data sourcing practices for model training, clarifying previously implicit data collection methods.
New explicit commitment regarding children's data protection and age restrictions, formalized as a distinct policy provision.
New detailed procedures for exercising privacy rights including appeal mechanisms and non-discrimination assurances, previously not explicitly documented.
New explicit commitment not to sell personal data and new opt-out mechanism for targeted advertising, addressing emerging privacy law requirements.
This provision was retained but severity was downgraded from medium to low; the practice of storing full conversations via feedback mechanism remains but is now considered lower-severity.
Removal of explicit third-party service integration disclosure may reduce transparency about data flows to external services.
Removal of explicit law enforcement disclosure provision reduces transparency about circumstances under which personal data may be shared with authorities.
Severity downgraded from high to medium, and provision name changed to reflect 'carve-out' language instead of 'override,' though the substantive policy text remains identical.
Severity downgraded from high to medium, and provision name changed from 'Exclusion' to 'Limitation,' though the substantive policy text remains identical.
Severity downgraded from medium to low, and the provision was narrowed to focus only on conversation deletion (removing the general deletion rights language).
Severity downgraded from medium to low, and provision name changed from 'Disclosure' to 'Transfer,' though the substantive policy text remains identical.
2 provisions unchanged.
View full change record →Monitoring
Anthropic has updated this document before.
Monitor includes same-day alerts, structured change summaries, and monitoring for up to 25 platforms.
Compliance Governance Intelligence
Need provision-level monitoring and regulatory mapping?
Compliance includes governance timelines, compliance memos, audit-ready analysis, and full provision tracking.
Start Compliance free trialCross-platform context
See how other platforms handle AI Model Training Data Use and Opt-Out and similar clauses.
Compare across platforms →Anthropic is more transparent than most AI companies about data retention. Here's exactly what happens when you delete your data, and how t…
Governance Monitoring
Structured alerts for policy changes, governance events, and provision updates across 318+ platforms.