Internet search engines are called to reinvent themselves, and it seems that Brave Search has decided to act quickly in this regard.. As you already know, Microsoft introduced the new Bing a few weeks ago and, since then, every few days we have heard about improvements and new features in its revolutionary chatbot. Google, for its part, announced Bard, whose beta tests should begin this March, but whose scope is still unknown, since those of the search engine are being particularly cautious after the initial announcement.
It is to be expected that all search engines, sooner or later, will move in this direction, integrating functions based on artificial intelligence to improve the user experience. Now, we can be sure that not everyone will opt for the same model. It is true that due to the success of ChatGPT, suddenly it seems that everything happens through the chatbots but, fortunately, we have already seen other equally interesting models but with a different way of working, such as the summary function that we will see soon in Microsoft Edge and also in Opera.
And that is precisely the case of Brave, the search engine born in the shadow of the browser of the same name and which, like it, focuses its proposal on the user’s privacy. Although there had been no previews or leaks previously (something very consistent with their philosophy, I must say), Brave has announced Summarizer, an artificial intelligence tool to get relevant and synthesized results in Brave Search.
The new function, which It is now available in the search engine, responds to user searches using a Long Language Model (LLM), which analyzes the results of the same, generates a response exclusively based on them and, in addition, highlights the key elements in the excerpts of the results displayed in the list.
Here, therefore, we find a very important difference with respect to ChatGPT and other generative models, since the answers provided by Brave Search with Summarizer are based exclusively on the search results, not on the information previously processed during their learningand even less in inventions like the ones we can frequently see in the OpenAI chatbot.
Like the new Bing, Brave Search provides, along with the answers, the list of fonts used for their preparation, which allows the user to confirm that the information in the summary is correct. In any case, since the type of model used is not generative in the purest sense of the concept, the risk of invented and wrong answers is substantially reduced.
