Google has introduced a new personalization feature for its Gemini chatbot, designed to improve response quality by leveraging users' search history. This feature is powered by the Gemini 2.0 Flash Thinking Experimental model, which only references search history when the AI determines it would be "helpful" for crafting a response.

For instance, when users ask Gemini for restaurant or travel recommendations, the chatbot can refer to their recent food-related searches to offer more tailored suggestions. This is part of Google's broader push toward personalization, with future plans to connect Gemini with other apps like YouTube and Google Photos. This integration aims to deliver "more personalized insights based on a deeper understanding of users' activities and preferences."

Users have the option to disconnect Gemini from their search history at any time. When receiving responses, users will see a summary of how Gemini arrived at its answer, including whether it referenced saved information, past conversations, or search history. A prominent banner will also display a link to disconnect from search history.

Currently, Gemini and Gemini Advanced subscribers on the web can enable this feature by selecting "Personalization (Experimental)" from the model dropdown menu. The feature is being rolled out progressively to mobile users and is already available in over 40 languages across "most" countries.

In addition, Google announced several other updates. These include allowing all Gemini users to create their own custom AI assistants, called Gems, for free. The company also revealed that the Gemini 2.0 Flash Thinking Experimental model will be integrated into its Deep Research feature, which is expected to enhance the chatbot's capabilities "across all stages of research."

Gemini's integration with Calendar, Notes, Tasks, and Photos will also be upgraded to the Gemini 2.0 Flash Thinking Experimental model, following earlier updates to YouTube, Search, and Google Maps.