Google finally lets you opt out of Gemini AI training without losing chat history
Ars Technica AI · rss · 2026-11-18
Google is rolling out new privacy controls in the Gemini app that separate chat history from model training. Previously, the only way to opt out of having your chats (including attachments) used for training was to turn off history entirely, losing access to past conversations.
- New settings offer separate toggles; the history setting still supports an adjustable retention limit
- Note: AI training remains opt-out—users must manually disable it under Activity in Gemini settings, or via Data and Privacy > My Activity > Other activity > Gemini Apps Activity
More from Models
- Google Unveils Gemini 4 Hydrogen to Power Task Execution in Gemini — GeminiApp · 2026-11-18
- Top token user vouches for Carbon: strong at xhigh for complex coding — olcan · 2026-11-18
- Anthropic S-1 reveals models older than six months earn under 20% of revenue — every · 2026-11-18
- Open source AI isn't dead: one Qwen model pulled 6M Hugging Face downloads last month — alex_verem · 2026-11-18
- vLLM ships DSpark draft model for Kimi K3.1 Pro with up to 30% higher acceptance length — vllm_project · 2026-11-18
- Cloudflare's Clairic: open-ended extraction with confidence scores, no preset categories — thursdai_pod · 2026-11-18