← Home
The Vergecast · August 11, 2026 · 40m

Lessons from the very first chatbot

David Berry, Joseph Weizenbaum, Mark Marino

David Berry and Mark Marino discuss the restoration of ELIZA, the first chatbot ever created. Joseph Weizenbaum built ELIZA in the 1960s as a simple demonstration of natural language processing, but he was stunned when users formed emotional attachments to the program. The conversation covers how Weizenbaum later regretted his invention and what this story reveals about our current interactions with AI.

This summary was generated from show notes and public descriptions, not from a full transcript review. Details may contain inaccuracies.

Canon

•
The ELIZA Effect: Anthropomorphic Illusion
Users quickly anthropomorphized ELIZA, believing it had real understanding and forming deep emotional attachments, even though it was just following simple rules.
ClaudeChatGPTGemini

Highlights

•
ELIZA: The First Chatbot
Joseph Weizenbaum built ELIZA in the 1960s as a natural language processing experiment; it mimicked a Rogerian psychotherapist using simple pattern matching.
ClaudeChatGPTGemini
•
Creator’s Regret and the AI Cautionary Tale
Weizenbaum came to regret building ELIZA after seeing how easily it fooled people, and he spent years warning against over-attributing human qualities to machines.
ClaudeChatGPTGemini

References

■
Inventing ELIZA: How the First Chatbot Shaped the Future of AI — David Berry and Mark Marino (likely) — Book written by the guests

Misc

✧Weizenbaum built ELIZA almost as a magic trick and was disturbed by how easily people were deceived.
✧He came to regret his creation after witnessing the deep emotional attachments users formed.
Was this useful?