Google Gemini trains on paid subscribers' chats by default

Google's own disclosure says saved Gemini chats from personal accounts, including paid subscribers, are fed into training data by default. The only toggle that stops that behavior also wipes conversations after 72 hours. The policy drew fresh attention from a Hacker News post pointing at Google's own support page.

The relevant line from Google: Gemini Apps activity 'is used to improve Google services for everyone, including by training generative AI models, if your Keep Activity setting is on.' Keep Activity defaults to on for adults. Turning it off stops future chats from being used to improve Google AI, but as Notebookcheck summarizes Google's terms, the company still holds conversations for 72 hours, and the same setting deletes stored chats on that 72-hour window, eliminating long-term memory for the assistant.

A subset of chats also goes to human reviewers. Google's warning on the support page is blunt: 'Please don't enter confidential information that you wouldn't want a reviewer to see.' Those reviewed copies are 'retained for up to three years' even after a user deletes their activity, Notebookcheck notes, citing Google's own terms. The commercial Gemini API, Vertex AI and Workspace tiers sit under separate data terms. The complaint lands the same week Google flipped Gemini to opt-in for schools after backlash over consumer-style data handling defaults.

The grievance on Hacker News is that none of this is surfaced at checkout for paid personal subscribers, and that the only toggle that stops training is the one that neutralizes the product.