Google’s Knowledge Graph – the reviews are mixed

Back in May of this year, without much hoopla, Google debuted a new semantic web feature called the Knowledge Graph.  The Knowledge Graph warrants the attention of educators and librarians because it marks a departure from search results based on keyword hits to results based on word meaning (a little like the service librarians offer).

When it works it’s a nice, useful addition to regular search results that can be used to help the search distinguish between topics with similar names or terminology, such as Adrian the designer and Adrian the town in Michigan.  That being said, don’t expect it to appear with every search.  My own simple test searches on Edith Head and Jean Desses brought back a nice Knowledge Graph on the former, but just regular results for the latter.

When it doesn’t work, the results can be a bit laughable as well as seriously misleading – more misleading than poor search results because the information is presented in the authoritative form of a graph.  A recent post on ACRLog examines this issue in depth using the Knowledge Graph on the poet Walt Whitman as an example (it seems he recorded a gospel music album . . . Not!).

Many in the mainstream news media have heralded the Knowledge Graph as a step toward a more intelligent web.  (See the New York Times and New Yorker articles below). Others, such as news outlets focused on SEO marketing, technology, and privacy issues are more somewhat more critical. (See the Search Engine Watch, Gartner, and CNet reviews below).

As for myself, I always get excited about attempts at semantic web searching.  However, I would agree with many of the reviewers I refer to that at this stage in the game, if you’re going to search the web for this type of information, you’re better off at Wikipedia.