ELIZA Effects: Pygmalion and the Early Development of Artificial Intelligence

abstract:Since the 1970s, scientists and psychologists have described what they call ELIZA effects: the tendency to attribute intelligence to responsive computers. ELIZA effects can be traced back to a natural language processing program or chatbot named ELIZA and created by Joseph Weizenbaum at MIT in the mid-1960s and to Weizenbaum's inspiration, Eliza Doolittle, in Bernard Shaw's play Pygmalion. This article examines Shaw's influence on ELIZA and his anticipatory engagement with questions about machine "thinking" and the origin and nature of intelligence that continue to drive research in computer science. It illuminates Shaw's significance for Weizenbaum and the pathbreaking English mathematician and theorist of artificial intelligence Alan Turing. It also proposes how we might reread Shaw's plays, and Pygmalion in particular, in light of what Weizenbaum and Turing found in Shaw.