Semantic Web—Another run at Artificial Intelligence?

Another emerging trend underway, and for some time now, is the Semantic Web. In a nutshell it’s the next generation of web technology which will use semantics or “knowledge” of the person making a request to generate a response. To date, web searches are based on keyword matches, but in the future it will take much more into account such as the background of the requester, the information requested, and the context in which the request is made.

In many posts on the semantic web I find a connection to “artificial intelligence.” Having worked on artificial intelligence in the 1980’s I know something about the challenges of AI. First, AI suffered from a PR problem. If it was solvable, it was no longer AI. Second, AI was very particular about what kinds of problems it would solve – so particular that it had limited usefulness coupled with unlimited expectations from the users. In this post the proponents of the semantic web try to distance themselves from AI. Some argue that the Semantic Web pursues the current goals of AI which is to create machines that exhibit intelligent behavior rather than the original goals of AI which was to create human-level intelligence.

Social networking currently drives the development of the internet through social linking. The semantic web in its academic form drives knowledge through tagging information and linking together those tagged items. In this paper, the authors argue that Metcalfe’s Law (the value of the network increases exponentially based on the number of nodes in the network) will apply only if social networking and semantic tagging come together.

Best regards,
Hall T.