When OpenAI announced ChatGPT soon se vio cómo este conversational AI model could pose a threat to Google’s traditional search engine. The company even declared a “code red” to search for solutions, and months later launched its own chatbot, called Google Bard. Now the search giant wants to go further with a new and ambitious plan. One with an important problem.
Project Magi. This is the name of Google’s new initiative that involves more than 160 employees of the company and that has the objective of revolutionizing its search engine and planting a much more ambitious alternative to ChatGPT. This is what sources close to the project have indicated in The New York Times, where they reveal how the movement is being given to an increasingly worrying threat: those who used the Google search engine on their devices are no longer doing it.
Samsung threatens to pass to Bing. A few weeks ago, Google employees found out that the South Korean electronics giant was planning to replace Google with Bing as a search engine for a defect in their devices. That would mean not counting about 3,000 million annual revenues that come through this agreement, similar to what for example Google maintains with Apple -although much more important, if they estimate 20,000 million-, and which is even more relevant in its book of accounting According to internal messages to those who had access to the NYT, in Google the reaction to such a possibility was “panic”.
IA in the browser (and in more sites). The objective with Project Magi is to integrate artificial intelligence models into the search experience to achieve more personalized functions that try to anticipate the needs of the user. Lara Levin, spokeswoman for Google, explained that “No todas las sessiones de brainstorming o product ideas lead to a launch, but as we have said before, we are excited about the incorporation of new search functions based on artificial intelligence, and soon we will give more details about it”.
A search engine that learns how to use it. Among the improvements would be the continuous learning of the search engine based on our use of it. For example, it could offer lists of preselected options of products to buy, or also of information that we could be looking for, always with the intention of getting ahead of what we are asking for. The search engine would be more conversational, although it is not clear if there will be a traditional search and then the conversation option, as Bing now offers with ChatGPT.
He is also preparing his own Midjourney. Another of the products that can form part of this set of tools with IA would be GIFI, a generative IA of images that would be integrated into the Google image search engine. There would be other options, such as the integration of these functions in Google Earth or in a potential tool called Tivoli Tutor to teach how to speak a language using conversations generated by AI.
A (very) cautious launch. According to data from the NYT, the company hopes to offer the tools of Project Magi next month —probably taking advantage of Google I/O?—and add more functions in autumn. The initial deployment will be slow, and it is hoped that only a million people will have access to these options. Before a year it was planned that Project Magi would be available for 30 million users, but exclusively in the United States.
Meanwhile, ChatGPT and Bing continue their conquest. Ese is precisely from problematic. Google Bard has been available for a few weeks in the United States, although as we have seen, it can be used with a VPN from outside the country. Even so, Google is still going with lead feet and its launches are being especially timid, something understandable considering that a disruption of this dimension threatens its advertising business. But all that caution with Bard and with Project Magi is counterproductive, especially because Microsoft does not stop offering more and more services and solutions based on ChatGPT, and OpenAI has not stepped on the brakes even though some have asked it to do so. .
Image: Focal Photo
In Xataka | From another transcendental debate of the IA: “pause” its development before it goes too fast