Launched by Google in 2012 and described at the time as an attempt to move from strings to things — from matching the letters a user types to understanding the entity those letters refer to. It is, in architectural terms, a vast relational database of entities and the connections between them: people, places, organisations, concepts, and creative works, each with defined attributes and documented relationships to other entities in the graph.

The Knowledge Graph is the infrastructure through which Google understands, rather than merely indexes, the web. When a search engine encounters a term, it does not simply locate pages containing that term — it attempts to resolve the term to an entity in the graph, retrieve what it knows about that entity, and return results that address the underlying meaning of the query rather than its literal surface. This is the mechanism behind the information panel that appears alongside search results for a recognisable entity: the graph has resolved the query to a known thing and is displaying what it holds about that thing.

The implication for publishers who understand this — and the number who do is smaller than the literature on the subject might suggest — is considerable. A term that exists nowhere in the indexed web has no entity record. It is, from the graph’s perspective, a void. The first content to define, use, and contextualise that term establishes its entity record. Every subsequent appearance of the term across the web references, however indirectly, the originating source. The Knowledge Graph resolves the term to its point of origin. That origin becomes the authority.

This is not a loophole. It is the system working exactly as designed.