As a blizzard begins to blow outside, the Massachusetts Institute of Technology’s (MIT) Special Collections archives are silent. It’s as if silence accumulates with the falling snow. I am the only researcher in the archive, but there is a voice I am trying to hear.
I’m looking for someone; Let’s call her the missing secretary. He played a very important role in the history of computers, but he was never named. I’m at MIT as part of my research on the history of talking machines. You may know these as “chatbots”; computer programs and interfaces that use dialogue as the primary means of interaction between humans and machines. Maybe you’ve talked to Alexa, Siri or ChatGPT.
Despite the current furor over generative artificial intelligence (AI), talking machines have a long history. In 1950, computer pioneer Alan Turing proposed machine intelligence testing. The test asks whether a human can distinguish between a computer and a person through speech. Turing’s test spurred research in artificial intelligence and the emerging field of computing. We now live in the future he envisioned: we talk to machines.
I wonder why the early computer pioneers dreamed of talking to computers and what was at stake in this idea. What does this mean for the way we understand computer technology and human-machine interaction today? I find myself at MIT in the middle of this blizzard because it was the birthplace of Eliza, the mother of all robots.
Eliza’s speech
Eliza was a computer program developed by moustachioed MIT electrical engineering professor Joseph Weizenbaum in the 1960s. He aimed to make communication between humans and computers possible through Eliza.
Eliza took typed messages from the user, parsed them for keyword triggers, and used transformation rules (where the meaning of a phrase can be inferred from one or more other phrases) to generate a response. In its most famous version, Eliza was claimed to be a psychotherapist, an expert who responded to the user’s needs. “Please tell me your problem,” was the opening prompt. Eliza not only received input in the form of natural language, but also created the “illusion of understanding.”
The program’s name was a reference to the protagonist of George Bernard Shaw’s play Pygmalion (1912), in which a Cockney flower seller is taught to speak “ladylike”. Like the Audrey Hepburn musical in 1964, this Eliza took the world by storm. Newspapers and magazines hailed the realization of Turing’s dream.
Even Playboy played with it. Eliza’s legacy is significant. Siri and Alexa are direct descendants of this program.
Eliza’s narratives tend to focus on a Frankenstein-esque story about the inventor rejecting his own work. Weizenbaum was horrified that users could be “tricked” by simple software. He renounced Eliza and the entire “Artificial Intelligentsia” in the coming years, much to the chagrin of his colleagues.
But I’m not in the archive to hear Eliza’s or Weizenbaum’s voice. In all these accounts of Eliza, one woman appears again and again: our missing secretary.
missing secretary
In his accounts about Eliza, Weizenbaum constantly worries about one particular user:
My secretary watched me work on this program for a long time. One day he asked to be allowed to talk to the system. Of course he knew he was talking to a machine. But after following his type for a few sentences, he turned to me and said: ‘Would you please leave the room?’
Weizenbaum saw his answer as alarming evidence that “Extremely brief exposure to a relatively simple computer program can cause strong delusional thoughts in fairly normal people.” His reaction sowed the seeds of his later hatred towards his creation.
So who was this “fairly normal” person? So what did he think of Eliza? If the missing secretary played such an important role, why don’t we hear from her? In this chapter of the history of talking machines we consider only one aspect of speech.
Back in the archive, I want to see if I can recover the secretary’s voice to see what we can learn from Eliza’s user. I’m trying to find my way among Weizenbaum’s yellowed papers. Surely there will be evidence among the transcripts, code prints, letters and notebooks? There are some clues, such as the mention of a secretary in letters to and from Weizenbaum. But there is no name.
I extend my research to administrative records. I look through department papers and the collections of Project MAC, the hallowed center of computer innovation at MIT, Weizenbaum’s workplace. Unlucky. I contact the HR office and MIT’s alumni group. I have the patience of archivists who are always generous. When my last day comes, I still hear only silence.
listening to silences
But this hunt revealed some things. How few organizations have historically cared about the people who produce, organize and store most of their information.
In the history of institutions like MIT and computing more generally, the authors of these records (mostly low-paid, low-status women) remain largely unwritten. Our silent secretary is the perfect eraser, anonymous transcriber of the documents on which history is built.
The contributions of talking machine users (their labor, expertise, perspectives, creativity) are often ignored. When the model is “conversation,” it is easy to think of these contributions as effortless or unimportant. But downplaying these contributions has real consequences, not only for the talking machine technology we design, but also for the way we value human input in these systems.
With generative AI, we talk about user input in terms of “chats” and “prompts.” So what kind of legal status can “speak”? For example, can we claim copyright regarding these statements? What about the studies in which these systems are trained? How can we recognize these contributions?
The snowstorm is getting worse. In the announcement, it was announced that the campus would close early due to weather conditions. The voice of the missing secretary still haunts me. For now, the history of talking machines remains one-sided. A silence that haunted me as I walked home through the muffled, snow-covered streets.
Looking for something good? Cut through the noise with a hand-picked selection of the latest releases, live events and exhibitions delivered straight to your inbox on Fridays, every two weeks. Sign up here.
This article is republished from The Conversation under a Creative Commons license. Read the original article.
Rebecca Roach’s research was supported by a Leverhulme Trust Research Fellowship for the project “Machine Speech: Literature, Computers, and Speech after 1945” and facilitated by the expertise and patience of staff at MIT’s Distinctive Collections.