In its pursuit of turning its artificial intelligence chatbot into a personal assistant, Google (GOOGL) is tapping into what it does best — search.
The tech firm said Thursday, March 13, that it is launching an experimental version of Gemini with personalization, which can connect with other Google apps and services to “provide responses that are uniquely insightful and directly address your needs.”
This means Gemini can now use search history to inform its responses, Google said, adding that the chatbot will get access to other platforms such as YouTube and Photos in the coming months. Users have to give Gemini permission to reference history from Search, and can disconnect it anytime.
“We’ll only use your Search history when our advanced reasoning models determine that it’s actually helpful,” Google said, adding that it’s continuing to get feedback from early testers.
The personalization feature can be found through Gemini’s model drop-down menu. After prompting the chatbot, it “will analyze it and determine if your Search history can enhance the response,” Google said.
Gemini and Gemini Advanced subscribers can access the experimental personalization feature via the web starting on Thursday, and it is rolling out on mobile, too. Gemini with personalization will be available in more than 40 languages across most countries where Google operates.
The chatbot is powered by Google’s Gemini 2.0 Flash Thinking Experimental model, which it launched in its Gemini app last month. The model is trained to “strengthen its reasoning capabilities” by breaking down prompts step-by-step and showing users its “thought process” to see how it came to its response. In December, Google introduced Gemini 2.0, which “will enable us to build new AI agents that bring us closer to our vision of a universal assistant,” Google chief executive Sundar Pichai said.
Google also said on Thursday that it was giving the 2.0 Flash Thinking Experimental model a longer context window to process larger amounts of information, and new features such as uploading files. The company added that it is upgrading its Deep Research feature with the model. The feature, which Google is making available to regular Gemini users at no cost, can compile multi-page research reports in minutes after browsing hundreds of webpages.
The Gemini app is also getting Google’s Gems feature, which now allows all users to build their own personal AI expert on topics such as languages and math at no cost.
