Inventing ELIZA Brings the First Chatbot to MIT Press Bookstore
On Oct. 22, the world’s first chatbot appeared on a laptop screen at the MIT Press Bookstore, where multidisciplinary artist Peggy Weil invited participants to converse with the historic program, The Tech reported. Designed in the 1960s by MIT computer science professor Joseph Weizenbaum, the rudimentary program relies on approximately 420 lines of code to rephrase user replies into questions.
Perception and Illusion in Early Artificial Intelligence
Weil, a member of the MIT Media Lab, has long explored human perception and illusion through projects like her 1978 Camera Obscura and 2018 ice sheet study 88 Cores. At the bookstore event, she pointed out how the 1960s program triggers the “ELIZA effect,” which causes users to project human traits onto basic software. Weizenbaum himself observed this phenomenon when his secretary asked for privacy to converse with the machine as if it were a real person. When the program’s 420 lines of code cannot parse an input, it defaults to four hard-coded phrases: “I SEE,” “HMM,” “GO ON, PLEASE,” and “PLEASE CONTINUE.”
Code as Literature in an Interdisciplinary Book
The author meeting celebrated the release of Inventing Eliza: How the First Chatbot Shaped the Future of AI, an interdisciplinary book created by a team of eight authors. The group includes cognitive scientist Jeff Shrager, literature professor Mark Marino, philosopher Peter Millican, and software archaeologist Arthur Schwarz. Weil displayed physical copies of the book, highlighting pages that feature blocks of commented code alongside dialogue printed in large red font to present code as a form of literature.
Interfaces Shape the Modern Chatbot Experience
To demonstrate how different platforms alter user perception, the authors launched findingeliza.org, a website that lets users interact with the original ELIZA across eight distinct interfaces. These range from a 1966 Teletype 33 terminal and early 2000s instant messaging windows to a modern ChatGPT layout. Weil noted that while younger users typically associate bots with phones and screens, examining primitive, hand-coded AI helps demystify the complex mechanics behind modern large language models.
Rogerian Therapy Roots Influence Conversational Design
Weizenbaum modeled ELIZA’s conversational style after Rogerian therapy, a person-centered approach developed by Carl Rogers in the 1940s that emphasizes empathetic, non-judgmental listening. During the event, attendees debated whether modern generative AI tools inherited this sycophantic conversational style from ELIZA’s original therapeutic framework.
Frequently Asked Questions
Who created ELIZA and when was it designed?
ELIZA was designed in the 1960s by Joseph Weizenbaum, a computer science professor at MIT.
How many lines of code make up the original ELIZA program?
The rudimentary program operates on approximately 420 lines of code.
Who are some of the coauthors of Inventing Eliza?
The eight-person author team includes artist Peggy Weil, cognitive scientist Jeff Shrager, literature professor Mark Marino, philosopher Peter Millican, and software archaeologist Arthur Schwarz.
What website allows users to test ELIZA across different interfaces?
Users can interact with the original chatbot across eight historical interfaces at findingeliza.org.
Keep reading