Particle.news
Download on the App Store
6 ARTICLES
·
2h ago

Anthropic Asks Consumer Users to Decide on Training Claude With Their Chats, Adds Five-Year Data Retention

Opting in before the September 28 deadline activates training on future chats with five-year retention.

Image
Image
Image
Geometric modular network of interconnected abstract shapes and structural elements

Overview

  • Existing users are seeing an in-app prompt to choose, and new users set their preference during signup, with the pop-up’s default toggle set to on and acceptance taking effect immediately.
  • The change covers Claude Free, Pro, and Max accounts, including Claude Code on those accounts, and it does not apply to Claude for Work, Claude Gov, Claude for Education, or API use via services like Amazon Bedrock and Google Cloud’s Vertex AI.
  • Only new or resumed chats and coding sessions are eligible for training, while untouched earlier conversations remain excluded, and deleted chats are not used for future training.
  • Users can change the setting at any time; turning it off prevents future training use and keeps a 30‑day retention policy, but data already used in completed or ongoing training cannot be removed from models.
  • Anthropic says the policy supports model improvement and safety systems, and it claims to filter or obfuscate sensitive data and not sell user information.