Google is rumored to be using Gemini AI for something interesting: the project, codenamed Ellmann, will narrate or process a user's life. Some people find this scary.

We recently shared the news that Google's new AI, Gemini, is smarter than Bard, and one of its notable features is that it's multimodal, meaning it can process and understand non-textual information, including images, video, and audio. And now there are rumors that with his help they are working on an interesting and secret project in the Google witch's kitchen. What's new is a life story teller named – for now internally – Elman. The choice of the name is not a coincidence, as the idea was added by the writer and literary critic Richard David Ellman, who is famous for his comprehensive biography of the life and works of James Joyce.

See also  Russia strictly regulates how Netflix works

But what is this Elman project? Well, Ellmann analyzes photos, search queries, and other potentially stored information to identify important events, connections, and personal interests. It will use a large linguistic model, such as Google's Gemini model, to interpret data, uncover connections, and infer deeper meanings from seemingly disparate information.

The analysis will be led by a chat interface called “Ellmann Chat” that will allow users to ask questions about their life history and get completely personalized answers. Google envisions it as “ChatGPT that knows everything about a user's life.” It's an impressive prospect, if somewhat shocking, given the vast amount of data and personal information Google already knows about most Internet users. In addition to retrieving facts, Ellmann can summarize user preferences, such as favorite activities or food choices, based on data patterns and trends. All in all, rather than just answering questions and helping you do mundane things, Google's secret initiative aims to capture a “holistic view” of a user's life. You can also learn things about the user that aren't actually shared with them.

Do you find it scary? That I'm not alone in this. Securing AI access to such a large amount of personal data already raises serious concerns. Google claims that protecting users' data is the most important thing, but how do Elman's algorithms work? Can we influence what he learns and shares? There are no answers to these questions yet.

Of course, it's also true that Ellmann is a completely internal project at the moment, and Google hasn't confirmed whether it will ever make it publicly available.

See also  Disaster film Far from the World debuts on Netflix

If you want to know about similar things at other times, like it HVG Tech Department Facebook page.