This paper presents the method of automatic knowledge retrieval from the Web. In this approach, millions of Internet users who naturally and constantly create and update the content of the Web are treated as the "voluntary contributors" who provide knowledge concepts without any need for additional motivation or coordination of their efforts. Obviously, only a small part of the statements accessible on the Web are valid knowledge axioms. Below, the "web knowledge concepts" filtering and verification methods are described, based on the similarity measurements with the concepts found in the manually created knowledge database. The results obtained with the system demonstrated that it is able to automatically retrieve semantic equivalents of several concepts submitted by the volunteer contributors, as well as the concepts that while being valid entries to a knowledge database, provide more details compared to the ones found in the manually developed database.