Google wants gemini to get to know you better

In the Ai Chatbot wars, Google thinks that the key to keeping users is to serve content that they cannot get elsewhere, such as the answers shaped by their internet habits.

Thursday, the company announced Gemini With personalization, a new “experimental capacity” for its Gemini chatbot applications which allows GEMINI to draw Google applications and services to provide personalized responses. Gemini with personalization can use the activities and preferences of a user through the ecosystem of Google products to provide tailor -made responses to requests, according to the Director of Products Gemini Dave Lemon.

“These updates are all designed to ensure that Gemini looks less like a tool and more like a natural extension of you, anticipating your needs with really personalized assistance,” wrote Lemon in a blog article supplied with Techcrunch. “The first testers found Gemini with useful customization for brainstorming and obtaining personalized recommendations.”

Gemini with personalization, which fits into Google Search before expanding additional Google services such as Google Photos and Youtube in the coming months, arrives while Chatbot manufacturers, including Openai, try to differentiate their virtual assistants with unique and convincing features. OPENAI recently deployed The possibility for chatgpt on macOS to modify the code directly in the supported applications, while Amazon is preparing to launch a “Agent” reinimage of Alexa.

Lemon said that Gemini with personalization are fueled by Google’s experimental Gemini 2.0 Flash Thinking Experimental IA ModelA so-called “reasoning” model that can determine whether the personal data of a Google service, such as the user’s search history, are likely to “improve” an answer. Narrow questions informed by tastes and aversions, as “where should I go on vacation this summer?” And “What would you suggest that I learn as a new hobby?”

“For example, you can ask Gemini for restaurant recommendations and this will refer to your recent food related to food,” he said, “or ask for travel advice and Gemini will respond according to the destinations that you have already looked for.”

If all this looks like a nightmare of confidentiality, well, it could be. It is not difficult to imagine a scenario in which Gemini inadvertently disseminates the sensitive information of someone.

This is probably why Google makes Gemini with an opt -in personalization – and excluding users under the age of 18. Gemini will ask for permission before connecting to Google’s research history and other applications, said lemon, and will show which data sources have been used to customize the BOT responses.

“When you use the personalization experience, Gemini displays a clear banner with a link to easily disconnect your research history,” said Lemon. “Gemini will only include your search history when you have selected Gemini with personalization, when you have given permission of Gemini to connect to your search history and when you have Web activity and application on.”

Gemini with personalization will be deployed to Gemini users on the web (except Google Workspace and Google for Education Customers) from Thursday in the model drop -down menu of the application and “gradually” reach mobile after that. It will be available in more than 40 languages ​​in the “majority” of countries, said Lemon, excluding the European Economic Area, Switzerland and the United Kingdom

Lemon has indicated that the functionality may not be free forever.

“The limits of future use can apply,” he wrote in the blog post. “We will continue to collect user comments on the most useful applications of this capacity.”

New models, connectors and more

As additional incentives to stick to Gemini, Google has announced updated models, search capacities and application connectors for the platform.

Subscribers to Gemini Advanced, the $ 20 premium subscription per month from Google, can now use an autonomous version of 2.0 Flash Thinking Experimental which supports the attachments of files; Integrations with applications like Google Calendar, Notes and Tasks; and a context window of 1 million. “Context window” refers to the text that the model can consider at any time – 1 million tokens is equivalent to around 750,000 words.

Google said this latest version of 2.0 Flash Thinking Experimental is faster and more efficient than the model it replaces, and can better manage the prompts that involve several applications, such as “search for an easy cookie recipe on YouTube, add the ingredients to my shopping list and find grocery stores that are still open nearby.”

Perhaps in response to the pressure of Openai and its Newly launched tools For in -depth research, Google also improves Deep researchIts gemini function looking for the web to compile reports on a subject. Deep Research now exposes its stages of “reflection” and uses the experience of Flash 2.0 thought as a default model, which should lead to “better quality” reports which are more “detailed” and “insightful”, said Google.

Deep Research is now free to try for all Gemini users, and Google has increased the limits of use of Advanced Gemini customers.

Free users of Gemini also get JewelsCustomizable chatbots focused on Google’s subject in Gemini, which previously required an advanced Gemini subscription. And in the coming weeks, all Gemini users will be able to interact with Google Photos, for example, to search for photos of a recent trip, said Google.