What Was the First AI Chatbot? ELIZA to ChatGPT (Timeline)

The first AI chatbot was ELIZA, built by Joseph Weizenbaum at MIT and published in 1966. Rule-based bots such as PARRY (1972), ALICE (1995) and SmarterChild (2001) followed. ChatGPT, released by OpenAI on 30 November 2022, brought large language models that answer without a script.

Sixty years on, a chatbot no longer needs a script. Build an AI agent from your own website for free and it answers from your pages, not from a list of rules.

When did chatbots start?

Chatbots started in 1966 with ELIZA, which held a text conversation by spotting keywords and replying from a fixed script. Every old AI chatbot that followed, from PARRY to ALICE to SmarterChild, still ran on rules someone wrote by hand. The modern kind, which reads a company's own content and answers questions nobody scripted, arrived only with large language models.

This page traces that line in a dated timeline, sets ELIZA beside ChatGPT, explains why the old bots broke, and ends with what the history tells you if you want an AI agent answering questions on your own website.

Key takeaways

• ELIZA, in 1966, was the first program to hold a convincing text conversation.
• Old AI chatbots were rule based: keywords and replies written by a developer.
• Their weakness was always the same: no script, no answer.
• Large language models removed the script, so an agent can learn from a business's own website and documents.
• What took a research lab in 1966 now needs no code.

Who built ELIZA, the first AI chatbot?

Joseph Weizenbaum, a computer scientist at MIT, described ELIZA in the January 1966 issue of Communications of the ACM, in a paper titled "ELIZA: a computer program for the study of natural language communication between man and machine". The program looked for keywords in what you typed and answered with a pre-set phrase built around them. The technique was simple. The effect was not.

ELIZA's best-known script, DOCTOR, played a therapist, turning your own words back into questions. People confided in it as if someone were there. That gap between what a program does and what users believe it understands is still called the ELIZA effect, and it is worth remembering every time a chatbot sounds more capable than it is.

ELIZA proved a computer could hold the shape of a conversation. Everything since, including the AI agent on a business website today, has been an attempt to fill that shape with real understanding.

What is the chatbot timeline, from ELIZA to ChatGPT?

Six chatbots mark the turns in the story. Each one was famous in its day, and each one changed how a chatbot understood people or where it found its answers.

The names changed for decades. The method mostly did not, until large language models. The table below groups the same history by era. Read it by its last column: that is the moment a visitor on your website feels.

For a closer look at what happens inside a modern one, see how AI chatbots work, step by step.

How does ELIZA compare with ChatGPT?

The first chatbot and the one that made chatbots mainstream sit 56 years apart. Put side by side, the gap is less about cleverness and more about where the words come from.

That last row is the one that matters for a website. A general chatbot answers from what it learned in training. A business needs answers from its own pages, prices and policies, which is why an AI agent for customer service searches your content first and writes second.

What were old AI chatbots, and why did they break?

An old AI chatbot was a rule engine. A developer wrote keywords, decision trees and canned replies, and the program matched your message against those rules and picked one. That is why old chatbots felt like a phone menu in text form. Ask what the script anticipated and you got an answer. Phrase it slightly differently and you got "Sorry, I don't understand."

Three limits defined the whole generation:

• No real language understanding. Keyword matching breaks on typos, synonyms and any follow-up that depends on what was said a moment earlier.
• Every answer hand-written. The bot knew what its builders typed into it and nothing more, so coverage was always partial and maintenance never ended.
• Dead ends by design. When the script ran out, the conversation stopped. That is where the reputation for infuriating chatbots came from.

What actually changed with modern AI agents?

The practical difference is where the knowledge lives. You no longer write the answers. You point the agent at material you already have, and it grounds what it says in that. If you are new to the term, start with what an AI chatbot is.

Dante AI works this way: point it at a website, it learns that site's content, and one script tag puts an agent on the page that answers visitors from the site's own material. Our guide on how to train a chatbot on your own data covers what to feed it.

What does chatbot history mean for your website today?

Sixty years of chatbot history come down to one lesson for a business website: a bot is only as good as what it knows. ELIZA knew nothing, so it could only reflect your words back. The rule-based bots knew only what someone typed in. A modern AI agent knows what you give it, which means the work has moved from writing scripts to choosing the right material.

When we speak with new users, the hesitation is almost never technical. It is working out which pages and documents genuinely represent the business. Three steps get you there:

1. Point it at your website. Get started with Dante AI, paste your URL, and let it read the pages you already publish.
2. Add the documents your team keeps re-sending. Price lists, policies, FAQs, onboarding guides. The answers usually exist already, buried in email.
3. Paste one script tag into your site. The step-by-step version is in how to add AI chat to your website.

Not sure whether the chat you already run has left the rule-based era? The six checks below will tell you in a minute.

What should you expect once your agent is live?

• Visitors get an answer when they ask, not the next working day
• Repetitive questions stop reaching your inbox
• You can see, in writing, what people came to the site to find out
• When the agent does not know, it says so, and on plans with Human Handover a teammate can take the conversation

Give it one honest test before you decide anything. Write down the five questions your inbox answers most often this month, put them to the agent, and compare each reply with what you would have typed yourself. Weizenbaum's users were fooled because they could not see the script. You can: every answer traces back to a page or document you supplied, so when one comes back thin, you know what to add.

You can build and test the whole thing on the free plan, which includes 100 message credits a month with no card, and paid plans start at $40 per month when you need more volume.

Sources

• Joseph Weizenbaum, "ELIZA: a computer program for the study of natural language communication between man and machine", Communications of the ACM, January 1966.
• OpenAI, Introducing ChatGPT, 30 November 2022.
• Apple Newsroom, Apple Launches iPhone 4S, iOS 5 & iCloud, 4 October 2011.
• Computer History Museum, Chatbots Decoded (PARRY, ALICE).
• Wikipedia, List of chatbots (PARRY 1972, ALICE 1995, SmarterChild 2001).
• Dante AI pricing and Help Center, checked 25 October 2026.