Google updates Gemini: it will remember your preferences in Italia too

A series of updates are arriving in Italia that aim to turn Gemini into an assistant with cognitive continuity. The most relevant novelty for the Italia market is the introduction of the Memory function. Basically, the chatbot will no longer forget who we are once we close the browser window, but will start capitalising on the details, preferences and style that emerged in previous conversations to tailor responses. We are basically allowing Gemini to learn from our past conversations. An example? 'If you've previously used Gemini to generate ideas for a YouTube channel focused on Japanese culture,' the site says, 'you've used Gemini to generate ideas for a YouTube channel focused on Japanese culture. When you ask for "new ideas for content based on my interests," Gemini might suggest a video theme such as "My first time trying..." in which you film yourself trying a new Japanese food, a traditional craft (such as origami or tea ceremony) or even a new Japanese game." This option is also already present in ChatGPT and inside Claude although with different customisation rules.

The sense of memory

This evolution is not only a functional convenience, but represents a precise field choice in the relationship between man and machine. Google has decided to set the function as active by default, a sign that extreme customisation is now considered the necessary condition to make these tools truly indispensable in daily professional and private life. Of course, the possibility for the user to intervene on the settings or deactivate it remains, but the message is clear: the efficiency of the algorithm depends on its ability to get to know us deeply.

Big G has introduced tools designed to facilitate digital relocation from other competing artificial intelligences. This is an import function that allows memories, contexts and the entire chat history to be transferred directly to Gemini. In economic and market terms, this initiative aims to drastically reduce what analysts call switching cost, i.e. the effort and loss of data that usually discourages a user from switching service providers. Again, similar solutions can be found in the models of Antrophic and OpenAi.

By making it possible to import past experience from other apps, Google eliminates the frustration of having to start training its virtual assistant from scratch. It is an open challenge to the competition that focuses on contextual portability: anyone will be able to update Gemini based on what they have already built elsewhere, making the transition to Mountain View's ecosystem smooth and frictionless. The battle for dominance in the field of artificial intelligence will thus be played out not only on computing power, but on the ability to retain and enhance each individual's digital history.