Google has been investing greatly in machine learning and artificial intelligence for videos, speech, translation, and even search. For the past few months, Google recently shared that a large portion of the millions of queries it receives per second are being interpreted by an AI nicknamed "RankBrain." RankBrain turns written language into mathematical entities that it can understand. If it encounters a word or phrase it’s unfamiliar with, it can “guess” what other words or phrases have a similar meaning and filter the results accordingly. RankBrain is one of the hundreds of signals that contribute to an algorithm that determines what results appear on a Google search and how they are ranked. Since its operation, RankBrain has become the third-most important signal contributing to the result of a search query.
RankBrain was the product of 5 years of work for a team of 5 engineers at the company. This team grew to dozens once the AI system was launched across all of Google search early this year. Google periodically updates the system by feeding it loads of new data to help it better reason with new concepts.
Google’s efforts to launch AI into search indicates how companies are now entrusting portions of their primary business operations to machine intelligence. Facebook uses AI techniques to filter newsfeeds, and Microsoft is also using AI to further advance its Bing search engine.