Meta is working on a search engine that collects information online and provides it to users of Meta's AI chatbots. This work has been ongoing for at least eight months. In this way, Meta hopes to reduce its reliance on Google and Bing, no longer needing them to provide real-time information. On Monday morning, Meta's stock price hovered around a 0.3% increase, while Alphabet's A shares fell nearly 0.8% in the short term
Meta wants to build its own search engine so that their AI chatbots can crawl the internet for information and directly answer users' current affairs questions without always relying on Google Search and Microsoft's Bing.
According to The Information on Monday, October 28th, Eastern Time, a person who has spoken with the Meta search engine team said that the purpose of doing this is in case Google and Bing stop cooperating with Meta one day, Meta will have its own search engine to provide information to users. Currently, when Meta's AI users query news, sports, and stock information, they rely on data provided by Google and Bing.
Influenced by this news, in the early trading session, Google's Class A shares fell nearly 0.8%, while Meta initially rose over 1.9% and then hovered around a 0.3% increase.
Meta strives to reduce its dependence on other tech giants
Meta's founder, chairman, and CEO Mark Zuckerberg is working hard to reduce Meta's dependence on other tech giants. For example, Meta previously relied on Apple Inc., but due to changes in Apple's rules, it has become more difficult for Meta to earn advertising revenue through iPhone apps, which has been a headache for Zuckerberg.
To address this challenge, Meta recently chose to collaborate with Reuters to enhance its AI's ability to answer questions in current affairs and news-related fields, reducing its reliance on Google and Bing. However, it is not clear whether Meta will pay Google or Microsoft to help its chatbots answer questions. Zuckerberg mentioned in April this year that there is not much "financial interaction" between Meta and Google, but did not provide details.
Regardless of the current situation, Meta remains cautious about its future development, as more and more users are turning to AI applications like ChatGPT instead of traditional search engines like Google.
Although AI such as ChatGPT developed by OpenAI has powerful language understanding and generation capabilities, they do not have the ability to directly connect to the internet to obtain real-time information. Therefore, in some cases, they need to rely on external sources of information, such as search engines like Microsoft Bing, to get the latest information. Microsoft, as one of OpenAI's main investors, plays a crucial role in providing search functionality for OpenAI's AI.
Meta AI is Meta's core strategy, aiming to make people prefer using their chatbots and spend more time on Meta's various applications. Meta wants these chatbots to be able to do big things in the future, such as writing software on their own or helping businesses advertise. Meta prominently places these chatbots in its own applications, making it easy for users on Facebook, Instagram, WhatsApp, and Messenger to find and start conversations with them In August this year, Zuckerberg revealed that Meta AI has surpassed 185 million weekly active users, narrowing the gap with OpenAI's ChatGPT, which has 250 million weekly users. Users may have different purposes for using these two chatbots, with Meta's robot mainly used for social media, while ChatGPT is more commonly used by programmers for coding or code review.
The search engine team at Meta AI is led by senior engineering manager Xueyuan Su. The team has been working for at least eight months, responsible for accessing various websites and organizing web pages into a database, a process known as web indexing, similar to creating a directory for web pages. In the future, Meta's AI will be able to search this directory to answer user queries. The existence of this team was previously unknown to the public.
Meta's web crawling technology has been public since July. Meta stated that they use this crawler bot "to train AI models or improve products by directly indexing content online." However, some websites like The New York Times do not allow Meta's bots to access their content