Why People Share Secrets With ChatGPT, Explained by a Chatbot
▼ Summary
– ELIZA is historically considered the earliest chatbot, often noted for “fooling” a secretary who saw its creation by MIT professor Joseph Weizenbaum.
– The book *Inventing ELIZA* recovers the program’s original source code from the MIT Archives, revealing multiple versions and scripts beyond its famous “DOCTOR” persona.
– A reprinted dialog from ELIZA raises unanswered questions about whether the young woman was real or invented, and how the system generated its responses.
– The “ELIZA effect” describes people’s tendency to treat responsive computer programs as more intelligent than they are, projecting human complexity onto them.
– ELIZA’s power relates to Alan Turing’s “imitation game,” which challenged notions of machine thinking by having a man pretend to be a woman.
For over six decades, the story of ELIZA has been told and retold as the dawn of conversational AI. The narrative usually begins with a secretary, reportedly “fooled” by the program into believing it was a real psychotherapist. But that well-worn tale has always been missing a critical component: the actual source code. Our new book, Inventing ELIZA, pulls that original code from the MIT Archives, finally presenting a close reading of the program alongside previously uncovered dialogs for scripts beyond the famous “DOCTOR” persona.
What we discovered is that there was never just one ELIZA. There were many versions, each built with distinct technical innovations and designed to run a variety of scripts or personas. Inventing ELIZA aims to correct the historical record by confronting the misconceptions, multiple versions, and the long-missing code that defined the program. In this excerpt, we look at one of ELIZA’s earliest conversations, how it set the stage for decades of human-computer interaction, and how this seemingly simple program still speaks to the unbridled ambitions of today’s AI industry.
Men are all alike.
IN WHAT WAY
They’re always bugging us about something or other.
CAN YOU THINK OF A SPECIFIC EXAMPLE
Well, my boyfriend made me come here.
YOUR BOYFRIEND MADE YOU COME HERE
He says I’m depressed much of the time.
I AM SORRY TO HEAR YOU ARE DEPRESSED
That dialog has been reprinted endlessly, inspiring programmers and writers to imagine the chatbots that followed. Yet the closer you look, the more questions surface. Who was this young woman? Was she a real person or a creation of ELIZA’s inventor, Joseph Weizenbaum? How did the system generate its responses, and how much editing took place? Why did it work so well at drawing people in?
ELIZA, particularly through its “DOCTOR” persona, helped spark a specific mode of thinking and a deep anxiety about our relationships with machines. Weizenbaum explored this in his 1976 book, Computer Power and Human Reason, offering philosophical, social, and political critiques. He saw the unique interaction his program created as a sign that new forms of human-computer relations would have profound effects, and he attempted to both explore and challenge them. After witnessing the public’s reaction, Weizenbaum was startled by the quick, often emotional attachments people formed with ELIZA. He saw this as “clear evidence that people were conversing with the computer as if it were a person who could be appropriately and usefully addressed in intimate terms.” The tendency to attribute empathy and invest private feelings into a machine puzzled him. He was deeply concerned by how readily people equated rationality with computation and ascribed understanding and intelligence to systems where none existed.
This phenomenon later became known as the “ELIZA effect.” By 1991 the term appeared in online forums, but its use predated that by decades. Sociologist Sherry Turkle defines it as “our more general tendency to treat responsive computer programs as more intelligent than they really are. Very small amounts of interactivity cause us to project our own complexity onto the undeserving object.” Cognitive scientist Douglas Hofstadter describes it as “the susceptibility of people to read far more understanding than is warranted into strings of symbols,especially words,strung together by computers.” This description applies just as easily to today’s generative AI systems.
To grasp the power and provocation of ELIZA, we can look to the infamous challenge posed by computer scientist Alan Turing in his essay “Computing Machinery and Intelligence.” Turing asked, “Can Machines Think?” He premised his thought experiment on a parlor game about gender, not technology. A man and a woman are hidden in separate rooms. An interrogator tries to identify who is who by asking questions. The man tries to mislead, pretending to be a woman, while the woman tries to convince the interrogator of the “correct” answer. Both claim to be the “real” woman, challenging essentialist notions of gender. This framework, like ELIZA, reveals how quickly we project meaning onto systems that offer only the illusion of understanding.
(Source: Wired)
