Thanks to a blog post on ReadWriteWeb, there has recently been some scuttlebutt about Google dipping into the waters of semantic search. Marshall Kilpatrick writes:
"In what appears to us to be a new addition to many Google search results pages, queries about birth dates, family connections and other information are now being responded to with explicitly semantic structured information."
Following up on this blog post, Daniel Nations notes some pretty interesting results: Google claims Jesus born in 1963. Nations goes on to say:
"[...] it seems we are seeing a good example of the downside of the semantic web. Whether it is an algorithm trying to pull the information out of an unoptimized web page or semantic engine pulling the information from a categorized page, we run into the same basic problem that we run into with many Web 2.0 sites: Can we trust the information?"
While the issue of information trustworthiness is interesting, it's not a problem unique to semantic search. Similar issues arise in keyword search. How can you trust the keywords on a page? Web masters have been inserting misleading keywords into their web pages to "fool" keyword search engines for years. In fact, the entire field of search engine optimization is the art of gaming the system. The reason that the problem seems more acute in the case of semantic search is that we have higher standards and expect authoritative answers to our questions.
But that expectation is questionable. Semantic search isn't synonymous with question answering. In fact, semantic search doesn't have to answer questions at all. Why? Because a semantic search engine doesn't have to tell us definitively the answer to our questions. It only needs to find the web pages that most directly address the question asked, and then leave it to the user to decide whether the information provided is "true". This is where the social web meets the semantic web. Web 2.0, meet Web 3.0.