ELIZA Simulator: Revisiting the Origins of AI Chatbots

eliza-simulator is a web-based ELIZA chatbot simulator that uses pure HTML, CSS, and JavaScript to recreate the classic ELIZA program developed by MIT in the 1960s, demonstrating how early rule-based dialogue systems created the illusion of intelligent conversation.

• Original Author/Maintainer: manuka-rashen
• Source Platform: GitHub
• Original Title: eliza-simulator
• Original Link: https://github.com/manuka-rashen/eliza-simulator
• Publication Date: 2026-05-25

Today, as large language models like ChatGPT, Claude, and Gemini sweep the globe, we seem to have grown accustomed to having smooth, natural conversations with AI. But few people know that the exploration of human-machine dialogue dates back more than half a century. In 1966, Professor Joseph Weizenbaum of the Massachusetts Institute of Technology (MIT) developed a program called ELIZA, which is widely regarded as one of the earliest and most influential chatbots in history.

Now, developer manuka-rashen has created a web-based ELIZA simulator, allowing us to experience this piece of AI history using modern browsers. This project is not only a nostalgic technical recreation but also a vivid lesson in computer science history.

ELIZA was born in the 1960s at MIT's Artificial Intelligence Laboratory, when computer science was still in its infancy. Weizenbaum's original intention in developing ELIZA was to prove that computers could simulate human dialogue through simple rules. The program's name comes from Eliza Doolittle, the heroine of George Bernard Shaw's play Pygmalion, implying that computers could also learn to 'speak' like Eliza.

ELIZA's most famous version is DOCTOR, which simulates a Rogerian psychotherapist. This therapeutic approach emphasizes guiding patients to explore themselves through questions and restatements rather than giving direct advice. ELIZA's algorithm is exactly good at this dialogue mode: identifying keywords, converting them into rhetorical questions, and creating the illusion of 'I am listening'.

ELIZA is a milestone in the history of AI development:

• It proved that even very simple rule-based systems can create the illusion of intelligence
• It sparked profound philosophical discussions about 'whether machines can understand language'
• It inspired decades of subsequent natural language processing research
• It demonstrated the psychological dimension of human-computer interaction—people tend to attribute intelligence to systems that can converse fluently.

eliza-simulator is a pure front-end project implemented using HTML5, CSS3, and native JavaScript, without relying on any external libraries or frameworks. This technical choice itself is a tribute to the early computing era—back in the 1960s, programs had to be concise and efficient because computing resources were extremely limited.

The project adopts a 1990s Windows-style user interface, and this design choice is quite meaningful:

• Visual Nostalgia: Gray tones, pixelated feel, classic window borders, evoking memories of the early personal computer era
• Functional Simplicity: No over-design of modern web applications, focusing on core dialogue functions
• Cross-Era Comparison: Allowing users to experience the technological evolution from 1960s algorithms to 1990s interfaces and then to modern browsers.