Anthropic Updates to Consumer Terms and Privacy Policy

Effective Date: August 28, 2025 Action Required by: October 8, 2025 (for existing users)

Key Changes & User Choice

Anthropic is updating its Consumer Terms and Privacy Policy to allow users to opt-in to having their data used to improve Claude models and safety systems.

• Purpose: To improve model capabilities (coding, analysis, reasoning) and strengthen safeguards against harmful usage (scams, abuse).
• User Control: Participation is optional. Users can change their preference at any time in their Privacy Settings.
• Scope: Applies to users on Claude Free, Pro, and Max plans, including use of Claude Code from these accounts.
• Exclusions: Does not apply to Commercial Terms services, including:
  • Claude for Work (Team/Enterprise)
  • Claude for Government
  • Claude for Education
  • API use (including via Amazon Bedrock and Google Cloud's Vertex AI)

Extended Data Retention

• If you opt-in: Data retention is extended to five years for new or resumed chats and coding sessions.
• If you do not opt-in: The existing 30-day data retention period continues.
• Important: Deleting a conversation or your account excludes that data from future model training. The new retention period also applies to user feedback.

Required User Actions

• Existing Users: Will see an in-app notification. You must accept the updated terms and make your training data selection by October 8, 2025 to continue using Claude.
• New Users: Can select their preference during the signup process.
• All Users: Can change their selection at any time in Privacy Settings.

Privacy Protections

"To protect users' privacy, we use a combination of tools and automated processes to filter or obfuscate sensitive data. We do not sell users' data to third parties."

FAQ Highlights

• What's Changing? Anthropic will train new models using data from opted-in Free, Pro, and Max accounts. The setting applies only to new or resumed chats and coding sessions.
• Why? Real-world interaction data provides valuable signals to improve model accuracy and usefulness, creating a feedback loop for improvement.
• Why Extend Retention? AI development cycles are long (18-24 months). Consistent data helps make model upgrades smoother and improves classifiers for detecting misuse over time.
• What if I Change My Mind? Turning off the setting stops future use of your data for training. Data already included in past or ongoing training runs will not be removed.