The Chatbot That Foretold Why People Share Secrets With ChatGPT

2026-07-14 · Source: WIRED - Ai · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Advanced, long

Summary

A new book, "Inventing ELIZA," recovers the original source code for ELIZA, the chatbot created by MIT professor Joseph Weizenbaum in the 1960s. This investigation reveals multiple program versions and scripts beyond its famous "DOCTOR" persona, correcting historical misconceptions. ELIZA's early interactions, such as the "boyfriend" dialog, demonstrated the "ELIZA effect"—the human tendency to project intelligence and empathy onto computer programs, a phenomenon Weizenbaum found concerning. The article connects ELIZA's design to Alan Turing's imitation game and the performance of identity, noting its namesake, Eliza Doolittle, highlights performative understanding rather than genuine intelligence. ELIZA's legacy extends to influencing natural language processing techniques and foreshadowing contemporary large language models like ChatGPT, particularly regarding the obfuscation of underlying machinery, the abstraction of human labor, and the ethical concerns of dehumanization and exploitation that Weizenbaum warned about.

Key takeaway

For AI scientists and ethicists developing or deploying conversational AI, recognize that the "ELIZA effect" persists, leading users to over-attribute intelligence and empathy. You must prioritize transparency about your system's actual capabilities and limitations to prevent user exploitation or dehumanization. Critically examine how your AI interfaces obfuscate underlying statistical models and human labor. Design systems that explicitly address ethical and social impacts, ensuring language is not abstracted from its human context.

Key insights

The "ELIZA effect" highlights humans' persistent tendency to project intelligence onto responsive computer programs, a critical lesson for modern AI.

Principles

• Human-computer interaction shapes user perception profoundly.
• AI systems can perform identity without true understanding.
• Obfuscation in AI risks exploitation and dehumanization.

In practice

• Examine AI interfaces for underlying statistical, rule-based, and human labor components.
• Design AI systems considering broader ethical and social impacts.
• Recover historical software artifacts to understand technological evolution.

Topics

• ELIZA Chatbot
• Human-Computer Interaction
• ELIZA Effect
• AI Ethics
• Large Language Models
• Natural Language Processing

Best for: AI Scientist, AI Ethicist, Research Scientist

Related on AIssential

• The First Chatbot’s Multiple Personalities — ELIZA's unearthed code reveals it was a sophisticated, multi-persona platform, not just a simple pattern-matching chatbot.
• AI is teaching us to speak like bots, and it’s a problem — AI chatbots are subtly homogenizing human language, making our communication increasingly machine-like by nudging expression towards an average.
• Apple’s new Siri AI knows when to shut up — Apple's Siri AI adopts a concise, utility-focused personality, contrasting with other chatbots' verbose and engaging styles to avoid user over-attachment.
• Someone, Not Something: How Language Defines How We Relate — English grammar's "it" pronoun for non-humans fosters objectification, hindering ethical human-AI relationships.
• Quoting Joseph Weizenbaum — Simple computer programs can induce powerful delusional thinking in normal individuals.
• ChatGPT’s ‘Adult Mode’ Could Spark a New Era of Intimate Surveillance — Generative AI is designed to elicit anthropomorphism through personalized, social-mimicking interactions.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by WIRED - Ai.