Anthropic's August 2025 consumer-policy announcement connected permission to train on conversations with longer data retention. For users who allowed model training, it announced five-year retention for new or resumed chats and coding sessions; users who declined would retain the existing 30-day period. The announcement covered Claude Free, Pro and Max accounts, including associated Claude Code use, while excluding services governed by commercial terms. It separately assigned five-year retention to submitted feedback. The training preference therefore affected both how eligible conversations could be used and how long they could be kept. Policy announcement.
