Google will add a dose of AI to Chrome, and this is what it looks like

Soon, finding a web page in Chrome’s history will be child’s play thanks to artificial intelligence. Google has in fact integrated a functionality allowing searches based on the real content of the sites visited, through simple queries in natural language.

Google’s improvements to its applications now focus on the integration of artificial intelligence. And the Chrome browser, although already very popular, is no exception to the rule. A valuable asset has just been spotted in the Canary version (the unstable avant-garde) of the browser: a “Historical search powered by AI” function within the dedicated section.

Concretely, it will be possible to carry out natural language queries to find the web pages consulted in the history, based on their real content, and not just their title or URL. This boon will change the way we navigate, but for the moment it is not without raising some concerns about confidentiality and compliance with the GDPR.

What this new AI-powered browsing history search feature does

This next-generation historical search promises to revolutionize the web browsing experience. No more painful scrolling through the mass of URLs and titles to find the content you want! With AI, just a few generic keywords will be enough for Chrome to link to relevant pages. It was X.com user Leopeva64 who noticed this new function and shared it with his subscribers on his account.

To give you a concrete example, let’s say that you were looking in your history for this site on which you finally found the words of this famous slow song which gave you your first romantic feelings. Don’t panic if the details escape you. A simple word from the chorus or lyrics, or just the term “slow” will allow the AI ​​to instantly find the corresponding pages in your recent history.

This feat is made possible by artificial intelligence technology, capable of grasping the context behind vague queries to deliver intelligent and intuitive results. In short, the AI ​​will search all pages for the keywords you typed. New filter options by date, category, etc. will also help refine the search.

The interface would show animations when the AI ​​“thinks” before proposing the best results. The AI ​​Gemini logo also suggests that Google uses this same AI model as for other of its services.

Although full technical details are still lacking, this novelty already promises to greatly facilitate the retrieval of information according to its substance, without depending on specific keywords. A welcome saving of time and simplicity.

Limited functionality on the unstable version of Chrome, Canary, which raises privacy and deployment questions

While this progress is something to be excited about, you will have to wait before taking full advantage of it. In fact, only the Canary version of Chrome, known to be unstable but a pioneer of new features, currently has this option, which is still under development.

The vast majority of users of the stable version will therefore have to be patient. A gradual deployment seems inevitable to refine the integration of AI and its diffusion on a larger scale.

Beyond this limited availability, the new text also raises thorny questions of data confidentiality. Although page content is locally encrypted, Google recognizes that its teams access queries and results to improve AI models. And what about GDPR compliance for European users? Clubic did not have access to this feature on Google Chrome Canary, which probably means that at the time of publication of this article, users affected by the General Data Protection Regulation do not yet have access to this functionality, and that Google is considering adapting it to European rules.

This is a worrying admission for the most cautious, which revives fears of increased surveillance of browsing habits under the guise of technological innovation. Google’s transparency contrasts with Copilot’s recent controversies, but the balance remains delicate between convenience and respect for privacy.

The analysis of this new document provides important clarifications. On the one hand, we better understand the key role of the Gemini AI model, already used by Google in other services, to power this improved historical search.
On the other hand, the text highlights Google’s efforts to improve the display and organization of history, with new options for grouping by date and a more descriptive presentation of the data collected.

However, these improvements appear to be still in development and only available to a few users of the Canary version. A beta deployment should be the next logical step.

Source : MSPowerUser