At Google, a team of computer scientists has projected a way to position the search results not by how famous the web pages are, but by their factual correctness.
To be clearer, this is 100% theoretical. It is a research paper, and not an announcement for a new product or anything that is equally exciting. Yet, the fact that a search engine could efficiently assess truth and that Google is vigorously pondering over that technology, should boggle the minds. After all, truth is a dicey, compliant thing and to grapple with it has conventionally been a solely human domain.
As per this recent research report, however, it is not too tricky for computers to find out if a given statement is true or false. Basically, to verify a stated fact, one only requires two things, that is, the fact and one reference work to evaluate it with. Google already has the initials of that reference work as the form of its knowledge graph, which shows August 15, 1990 if you search for the birthday of Jennifer Lawrence, or American when you search for Obama nationality.
Google pick those details mostly from services like Wikipedia, Freebase, and the CIA World Factbook. Knowledge Vault, an individual internal research database can also extract facts automatically from the text given on web pages. Whichever database one is talking about, Google arranges these little factoids as things, which is called the ‘knowledge triples’ - subject, attribute, relationship.