The term "knowledge acquisition bottleneck" has been used in Word Sense Disambiguation Tasks (WSDTs) to illustrate/express the problem of the lack of large tagged corpora. In this paper, an automated WSDT is based on text corpora extracted / collected from Internet web pages. First, the disambiguation for the sense of a word, in a context, is based on the use of its definition and the definitions of its direct hyponyms in the WordNet to form queries for searching the Internet. Then, the "sense-related examples", in other words the collected answers / information, are used to disambiguate the word's sense in the context. A (similarity) metric is used to calculate the similarity between the context and the "sense-related examples" and the word is assigned the sense of the most similar example with the context. Some experiments are briefly described and the evaluation of the proposed method is discussed.