Meta is working on an AI-powered search engine, a new report from The Information says. The search tool would provide automated summaries of current events and news to power the AI chatbot available through Facebook and Instagram.
Meta's chatbot relies on web searches from Google and Microsoft Bing to answer users' questions. However, a team at Meta has been working for more than eight months on an internal database that would enable the chatbot to retrieve information without going through a search engine outside of itself.
The source says Meta has taken steps to compile its own web index by deploying spiders to crawl and catalog internet pages. Collected data will be used in training machine learning models to return relevant results on queries about topics, people, and events.
This is both a huge opportunity and a huge challenge for Meta, as creating its search engine could make its chatbot more useful while keeping user data and interactions confined to its platforms. On the other hand, scraping web content also raises serious legal issues around copyright and fair compensation for creators.
Meta also announced a multi-year deal with Reuters last week. This deal will provide legal access to Meta's AI chatbot to source full news articles and content from Reuters.
Other giants, including Apple, OpenAI, and Perplex, are also making their moves into AI search. Software updates from Apple hinted that it could work on building search for its App Store. Perplexity has faced lawsuits from publishers such as News Corp and The New York Times over its methods.
OpenAI confirmed that it is working on a model called SearchGPT. "SearchGPT will quickly and directly respond to your questions with up-to-date information from the web while giving you clear links to relevant sources," OpenAI wrote in a blog post.
Source: The Information
5 Comments - Add comment