The book Inventing ELIZA recovers lost source code from MIT’s archives to reexamine the 1960s chatbot’s historical influence on modern artificial intelligence and human-computer interaction.

Key Points

• ELIZA was created by MIT professor Joseph Weizenbaum in the 1960s as a program capable of simulating a psychotherapist through scripted responses.
• The authors of Inventing ELIZA recovered the original source code from MIT Archives to analyze the program's multiple versions and personas.
• The "ELIZA effect" describes the human tendency to project intelligence and empathy onto simple computer programs, a phenomenon still relevant to modern generative AI.
• Weizenbaum’s work, particularly his 1976 book Computer Power and Human Reason, warned against the dehumanizing potential of treating computational systems as sentient entities.
• The book explores how ELIZA’s design anticipated contemporary debates regarding algorithmic bias, gender performativity, and the ethics of natural language processing.

Why it Matters

• Understanding ELIZA’s history provides critical context for the rapid development of modern large language models like ChatGPT, which often rely on similar chatbot interfaces to mask complex statistical processes. By analyzing these early systems, researchers can better identify the persistent risks of dehumanization, data exploitation, and the overestimation of machine intelligence in today's tech industry.