Textbook of AI

Abstract. Weizenbaum's book-length critique of the AI project from inside MIT's AI Lab. He had built ELIZA, a 1966 chatbot that simulated a Rogerian psychotherapist with about 200 lines of pattern matching, and was horrified to discover that his secretary and other observers attributed genuine understanding to the program. The book argues that there are decisions human beings should not delegate to machines, not because machines cannot be made competent at them, but because the act of delegation itself is corrosive to human responsibility. The book is one of the foundational works of AI ethics and remains widely cited in contemporary debates about chatbots, recommender systems and clinical AI.