Anthropic is changing its data policy: user chat transcripts and code sessions will now be used by default for AI model training. This also means user data will be retained for five years.
Key Changes:
- Applies to New Interactions: Only new chats and sessions are affected. Older interactions remain private unless re-engaged.
- Automatic Opt-In: Users are automatically included but can opt-out until September 28th.
- Consumer Subscriptions Affected: This impacts all consumer Anthropic subscriptions.
- Exclusions: Business Accounts: Claude Gov, Claude for Work, Claude for Education are unaffected.
- API Users: Users accessing via API or third parties are also exempt.
Anthropic states they use automated tools to filter sensitive data and commit to not selling user data. This policy aims to improve AI security and content moderation.
