Hi Gemini Product and Tech Team,
As a premium subscriber (Google One AI Premium), I am writing to strongly request a feature that is already standard across OpenAI, Anthropic, and xAI: The ability to turn off model training/improvements while keeping the sidebar conversation history active.
Currently, your “Gemini Apps Activity” forces a bundle between history retention and model training. For independent researchers and developers working on high-value proprietary projects, intellectual property, and sensitive system architectures, this is a massive privacy bottleneck.
Why must premium users choose between losing their ongoing project context or giving up their data privacy? This technical limitation does not exist on your competitors’ platforms, and it shouldn’t exist on Gemini Web either.
Please treat data privacy with the same priority as your AI capabilities. We need a standalone “Opt-out of model training” toggle that leaves our chat history intact.
Thank you.