HOT 285: Are My AI Chatbot Conversations Private? - The Risks Of Sharing Information with AI Chatbots
Summary
Host Micah Sargent answers a listener’s question about AI chatbot privacy, clarifying that conversations are stored on company servers and used for model training by default. He explains that while users can disable training, data remains accessible to human reviewers and is not covered by HIPAA or legal privilege. The episode details specific settings for ChatGPT and Gemini to manage data retention, highlighting the trade-offs like losing chat history in Gemini. Sargent advises using dedicated health tools or temporary chats for sensitive information and removing personal identifiers befor…
Topics discussed
Sponsor: Trusted Tech Microsoft 365 licensing savings
Listener question: AI chatbots vs. Google Search privacy
Short answer: Data storage and default training settings
How prompts, uploads, and voice are stored and used
Impact of turning off training on ChatGPT and Gemini
Human review, legal holds, and lack of privilege
Advertising, data retention, and the NYT lawsuit
Chatbot memory persistence and comparison to search
HIPAA limitations and the new ChatGPT Health tool
Advice for sanitizing health data before uploading
How to disable training and manage ads in ChatGPT
Managing activity and auto-delete settings in Gemini
Gemini temporary chats and the thumbs-up review trap
Verdict and quick-fix privacy settings summary
Listen ad-free on Castria