Technology

Claude AI Privacy Guide: Key Settings You Should Change

Claude AI Privacy Guide: Key Settings You Should Change
Table of Contents
  1. Stop AI Training on Chats and Voice Data
  2. Turn Off Location Metadata Collection
  3. Manage and Disable Chat Memory
  4. Use Incognito Mode for Sensitive Conversations

Users concerned about personal information and privacy while using the AI chatbot Claude can adjust several default settings to protect their data. While Claude offers strong privacy protections compared to competitors, changing specific configurations can limit or eliminate the amount of data the platform retains about you, as reported by CNET.

Stop AI Training on Chats and Voice Data

By default, most AI chatbots enable a setting that permits companies to train their models on user conversations. To disable this feature on web, iOS, and Android versions of Claude, users need to navigate to Settings, select Privacy, and toggle off both “Help improve our AI models” and “Allow us to use your voice data.” Enterprise accounts and similar work plans may not automatically collect this information, meaning some users might not see these options.

Users of Claude Code can also delete stored code sessions from Anthropic’s servers via this menu without removing local sessions from their computers. Additionally, any previously shared chats can be permanently deleted from this section.

Turn Off Location Metadata Collection

Anthropic uses device IP addresses to identify a user’s city and region to improve the product experience. On the web app, users can disable this by clicking their profile icon, opening Settings, selecting Privacy, and toggling off Location metadata.

For mobile applications and web extensions, location settings are managed through device systems. Android users can revoke permissions via phone settings under Apps, Claude, Permissions, and setting Location to “Don’t allow.” Alternatively, the Android Claude app features its own location setting that can be set to “Never.” On iOS and iPadOS, users can open the system Settings app, scroll down to Apps, and adjust Allow Location Access to “Never.” Disabling location may impact features requiring geographic data, though Anthropic notes that IP addresses are gathered to ensure terms compliance and prevent abuse.

Manage and Disable Chat Memory

Claude includes a built-in memory feature enabled by default, allowing the chatbot to recall personal preferences and context across different sessions. Users can ask the chatbot what information it remembers or view stored memories directly by going to the profile icon, selecting “Your Claude” in the settings sidebar, and clicking “Your memory.” From there, specific memories can be edited using a pencil icon or deleted entirely with a trash icon.

To stop Claude from generating new memories from ongoing chats, users can navigate to the “Your Claude” settings menu and toggle off “Generate memory from chats.”

Use Incognito Mode for Sensitive Conversations

Users who previously engaged in sensitive conversations while AI training was active cannot alter past data usage, but those chats can be found in the chat history and deleted manually. For future sensitive discussions, users should utilize incognito mode. Incognito chats are neither saved in the user’s history nor utilized for training AI models.

You May Also Like

Leave a Comment

Your email address will not be published. Required fields are marked *