Gemini Privacy Settings Update
3 ステップ1 件3 ソース
Google has updated the privacy settings for its Gemini app, decoupling chat history retention from generative AI model training. Previously, users had to choose between keeping their conversation history or opting out of having that data used to train Google's AI models. The new update allows users to retain their chat history while independently disabling AI training consent.
However, this change comes with a significant adjustment to data retention policies. For chats where the "History" setting is turned off (or temporary chats), Google will now store the data for approximately two months, up from the previous 72-hour limit. Google states this extended retention is necessary for safety and security purposes, including detecting abuse and responding to system failures, but explicitly confirms that these retained chats will not be used for model training.
This move aligns Gemini more closely with competitors like OpenAI’s ChatGPT and Anthropic’s Claude, which have long offered separate controls for history and training. While ChatGPT and Claude typically retain temporary chats for up to 30 days, Google’s new two-month window is longer, though justified by the company as essential for investigating complex emerging abuse scenarios. Users who previously disabled all activity tracking will find both new settings defaulted to "off," meaning their future chats will still be stored for two months for safety reasons, even if they are not used for training.
経緯
November 16, 2026: Google announces updates to Gemini app privacy controls, separating history retention from AI training opt-outs.
blog.google ↗November 17, 2026: TechCrunch reports on the extended retention period for temporary chats, citing a Google spokesperson's statement on safety needs.
techcrunch.com ↗