Google’s Gemini AI chatbot is about to get more personal. A new Gemini Personalization model discovered within the Google app beta (version 16.8.31) by Android Authority will allow Gemini to access your Google Search history to deliver more tailored responses.
This optional feature requires users to explicitly grant permission for Gemini to access their search history and necessitates enabling the “Web & App Activity” setting. Upon selecting the Gemini Personalization model, a confirmation pop-up appears, outlining the data sharing.
Gemini Personalization model.Gemini Personalization model opt-in screen. (Image credit: Android Authority)
It’s important to note that this is an experimental feature. Google has clarified that this data will not be used to improve Gemini itself, and users can revoke access at any time. Early tests conducted by Android Authority indicate that the personalized model provides accurate results based on search history, but they also caution about potential issues due to its experimental nature.
Like other large language models (LLMs), Gemini is trained on a massive dataset of text and code. This broad training allows it to leverage a vast knowledge base, including search results, to enhance its overall performance. However, this new personalization model marks a shift towards using individual search data to refine responses on a per-user basis.
Google hasn’t released further details about the Gemini Personalization model or its public availability. Speculation suggests it might launch alongside the scheduled update for Gemini Advanced users.
The introduction of personalized responses is bound to generate discussion around AI and user privacy. Whether this feature remains optional or becomes standard for Gemini will undoubtedly influence user adoption and the broader conversation surrounding AI ethics. It will be fascinating to observe user reactions and the long-term implications of this personalized approach to AI chatbots.
