The Web as a Platform for Experimental Human-Computer Linguistic Interaction

This chapter describes the use of the Web as a content generator for experimenting with human-machine language interaction. The main idea is exploiting the distributed and unsupervised complexity of the Internet in order toget all the semantic/syntactic material needed to carry on a linguistic text based interaction. After describing the main elements of the proposed Consciousnet architecture, the results of a set of Turing-inspired tests are shown to demonstrate how the unpredictability and generality of the environment can be used as a basis for designing tests and experiments involving both psychologists and AI scientists.