Birth of AI and Early Dialogues (1950s–1970s)

This essay is part of the Between Brain & Binary series. Read the overview and full reading order on the Between Brain & Binary page.

Introduction

By the mid-20th century, the dream of creating machines that could “think” was no longer a philosophical curiosity — it was a serious scientific pursuit. As psychology sought to model human cognition, computing scientists were asking a parallel question: could intelligence itself be simulated?

(L) Alan Turing (1912–1954) in 1936 at Princeton University. Photo attribution: Unknown photographer, Public domain, via Wikimedia Commons.
(R, top) Turing Machine – Turing’s First Machine. Attribution: Wvbailey, CC BY-SA 3.0, via Wikimedia Commons.

(R, bottom) Turing Machine Example – 3-State Busy Beaver. Attribution: Wvbailey, CC BY-SA 3.0, via Wikimedia Commons.

The Turing Test: Intelligence as Conversation

One of the most influential milestones came in 1950, when mathematician Alan Turing posed a radical question: Can machines think? (Turing, 1950). In his landmark paper Computing Machinery and Intelligence, Turing proposed a test — now known as the Turing Test — in which a machine would be considered intelligent if it could converse with a human so convincingly that the human could not reliably tell the difference.

This idea reframed intelligence from an internal property to an observable behaviour — something measurable through language and interaction. It also shifted psychology’s influence on computing: if cognition could be measured by how humans communicate, then machines, too, might “think” by mastering conversation.

(L) Joseph Weizenbaum (b. 1927), computer scientist and creator of the ELIZA software in 1966. Attribution:Ulrich Hansen, Germany (Journalist)., CC BY-SA 3.0 <http://creativecommons.org/licenses/by-sa/3.0/>, via Wikimedia Commons.
(R) ELIZA conversation interface example, where the program mimics a Rogerian psychotherapist via pattern matching. Image (R): screenshot from public archive / software archive, used under public domain or allowed license. File:ELIZA conversation.jpg, Public domain, via Wikimedia Commons.

ELIZA and the Birth of Digital Empathy

This concept was tested in practice just over a decade later. In 1966, computer scientist Joseph Weizenbaum created ELIZA, a simple program that mimicked a Rogerian psychotherapist by reflecting users’ statements back at them (Weizenbaum, 1966). Although ELIZA lacked any understanding of meaning, users often felt heard and emotionally engaged — a phenomenon Weizenbaum himself found unsettling. It revealed an enduring truth: people can project emotion and intention even onto lifeless code.

From Surface Mimicry to Inner Models: PARRY

In 1975, psychiatrist Kenneth Colby pushed the idea further with PARRY, a program designed to simulate the thought processes of a person with paranoid schizophrenia (Colby, 1975). Unlike ELIZA’s surface-level reflections, PARRY modelled internal states such as fear, suspicion, and hostility, producing conversations that felt more psychologically complex.

A famous demonstration involved a scripted “conversation” between ELIZA and PARRY, orchestrated by computer scientist Vint Cerf (1973). The dialogue, though artificial, demonstrated how even early AI systems could emulate therapeutic dialogue and emotional nuance — decades before neural networks and large language models would refine the craft.

Conversation excerpt between PARRY and ELIZA, two pioneering natural language programs developed in the 1960s and 1970s. This interaction illustrates early attempts to model therapeutic dialogue and emotional reasoning using symbolic Attribution: Public domain – available under Creative Commons Zero (CC0 1.0) license. Source: Cerf, 1973.

The Psychology of Simulation

These early experiments raised profound questions that still shape AI research today. Could a machine that mimics conversation be said to “understand”? Can a program that models emotion be considered empathic? And why do humans so readily attribute meaning and humanity to lines of text generated by algorithms?

ELIZA and PARRY were not just technological curiosities — they were psychological experiments in disguise. They showed that the boundary between communication and cognition, simulation and understanding, was far more porous than anyone had imagined. These programs marked the dawn of digital empathy — a concept that would define the next era of human–machine interaction.

This period marked a pivotal turning point — when machines first began to speak, mimic, and even simulate emotion. These early experiments reshaped our understanding of both cognition and communication, laying the groundwork for the digital empathy revolution that would follow.

Reference List (APA 7 format)

• Babbage, C. (1864). Passages from the life of a philosopher. Longman, Green, Longman, Roberts, & Green. https://www.gutenberg.org/ebooks/1216
• Cerf, V. G. (1973). Conversation between PARRY and ELIZA. Stanford AI Archives (Reproduced in multiple AI anthologies). Available via Creative Commons Zero (CC0 1.0). https://www.researchgate.net/figure/Conversations-between-PARRY-and-ELIZA-Cerf-1973_fig6_301874461
• Colby, K. M. (1975). Artificial paranoia: A computer simulation of paranoid processes. Pergamon Press. https://doi.org/10.1016/C2013-0-02410-3
• Dreyfus, H. L. (1972). What computers still can’t do: A critique of artificial reason. MIT Press. https://mitpress.mit.edu/9780262540671/what-computers-still-cant-do/
• Lovelace, A. (1961). Notes by the translator (original work published 1843). In B. V. Bowden (Ed.), Faster than thought: A symposium on digital computing machines (pp. 94–117). Pitman Publishing. https://www.fourmilab.ch/babbage/sketch.html
• Saygin, A. P., Cicekli, I., & Akman, V. (2000). Turing test: 50 years later. Minds and Machines, 10(4), 463–518. https://doi.org/10.1023/A:1011288000451
• Turing, A. M. (1950). Computing machinery and intelligence. Mind, 59(236), 433–460. https://doi.org/10.1093/mind/LIX.236.433
• Weizenbaum, J. (1966). ELIZA—a computer program for the study of natural language communication between man and machine. Communications of the ACM, 9(1), 36–45. https://doi.org/10.1145/365153.365168

Author Note (AI Usage):This article was drafted with assistance from a generative AI system to organize structure and suggest phrasing. All facts, interpretation, and final editing have been verified and approved by the author. The AI did not access any private health data. 

Continue in this series: Parallel Frontiers: Early Telepsychiatry and Remote Care (1950s–1970s). Or return to the Between Brain & Binary overview.