Your Claude chats are being used to train AI — here's how to opt out

Anthropic has revised its privacy policy to utilize user chat and coding session data for training future AI models, reversing a previous commitment to non-use. This default setting applies to all free and paid personal users, meaning their interactions may now contribute to model development unless they actively intervene. The company also extended data retention periods to five years, regardless of training participation choices, significantly increasing the window during which user data is stored and potentially accessible. This policy shift raises critical concerns regarding data ownership and informed consent in the AI industry. By making opt-out mechanisms the user’s responsibility rather than ensuring privacy by default, Anthropic prioritizes model improvement over individual control. Consequently, personal users must manually disable specific toggles to protect their information, while commercial and government accounts remain automatically exempt, creating a disparity in privacy protections based on account type and subscription level. Relevance to open data lies in the tension between proprietary data extraction and the principles of transparency and user sovereignty. When AI developers aggregate vast amounts of personal conversations without explicit, default consent, they challenge the notion that data usage should be open, predictable, and respectful of user autonomy. This move highlights the urgent need for clear standards in how AI entities handle data inputs, emphasizing that open practices must include respect for user privacy rights and clear, accessible opt-out pathways to maintain trust in the technology ecosystem.

Source: tomsguide.com
Published on 2025-10-02

Your Claude chats are being used to train AI — here's how to opt out

Related news