What’s in a name?
That which we call a rose
by any other name
would smell just as sweet
—William Shakespeare

We propose that a 2 month, 10 man (sic) study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire.

These are the first two sentences of a document titled, “A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence” dated August 31, 1955. It was written by John McCarthy, Marvin L. Minsky, Nathaniel Rochester, and Claude E. Shannon and is maybe the first documented use of the term Artificial Intelligence.

The term Artificial Intelligence was the brainchild of pioneering computer scientist John McCarthy, and the Dartmouth Conference of 1956 is often regarded as the founding event of AI as a field of study.

The conference was based on a powerful idea:

… the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it.

The rest, of course, is history.

What is less known however, is that Artificial Intelligence was NOT the first term that was considered for this nascent field of study, and in fact is a fascinating story that highlights the profound impact a name can have on the world.  

Note: I first heard of this story in an article by Amber Case titled: Inside the Very Human Origin of the Term “Artificial Intelligence” — And Its Seven Decade Boom/Bust Cycle. That led to some digging around on the Internet, and some introspection on my part – and this post.

But before we get into the history, please allow me to set the context with a brief digression on words and their ways.

A Digression on Words and Their Ways

Words matter.

Nowhere is this fact made more salient than in George Orwell’s novel “1984.” In this book, Orwell presents a dystopian world where language is a powerful tool of control, where the government systematically reduces and simplifies language to limit the range of ideas that can be expressed. Certain thoughts are, quite literally, unthinkable. For instance, the word “free” only exists in contexts like “This dog is free from lice,” but not in the sense of political or intellectual freedom. This creates an alternate reality where dissent becomes impossible because the very words to express it no longer exist. What Orwell recognized is that controlling language can shape perception, restrict critical thinking, and ultimately alter reality itself.

To be clear this power to shape reality through language is not limited to authoritarian governments. In our modern world, corporations, media outlets, interest groups, and even social movements wield significant influence over language and, by extension, public perception. Critical to this is the idea of “framing” i.e. the careful choice of words and phrases to evoke specific emotional responses or cognitive associations.

There are so many examples of this in the world around us—good and bad. Consider, for instance, energy companies rebranding fossil fuels as “natural gas” to emphasize their naturalness rather than their environmental impact. Military actions are softened by using “collateral damage” instead of “civilian casualties,” and controversial interrogation methods are masked by “enhanced interrogation techniques” rather than “torture.” On the other hand, the shift from “global warming” to “climate change” broadened the conversation around environmental issues and the move from “shell shock” to “PTSD” reflects a deeper understanding of trauma’s long-term effects. We see this in immigration debates where dialogue is framed as a choice between “illegal aliens” and “undocumented immigrants,” while the abortion controversy is influenced by terms like “pro-life” and “pro-choice”. And, of course, tech companies use terms like “sharing economy” to paint their business models in a positive light.

At some level this decentralized manipulation of language can be even more insidious than state control, as it’s often subtle, pervasive, and comes from sources we trust or identify with. As a result, our perception of issues, products, and even ourselves can be profoundly shaped by these linguistic choices, often without our conscious awareness.

In short, what we call something changes how we perceive it. Which brings us back to the origin of the term Artificial Intelligence.

Coming back to Artificial Intelligence

In the summer of 1955, as McCarthy was planning the Dartmouth conference, he had a choice to make about what he would call this nascent field. The initial and natural choice for naming this field was “cybernetics,” a term already in use to describe the study of control and communication in machines and living things. Cybernetics was a term coined by MIT scientist Norbert Wiener, from the Greek word for “steersman,” and he defined it as the “science of communication and control.” And therein lay a problem.

Norbert Wiener, at that time, was one of the most famous scientists of his era. A child prodigy (having graduated high school at age 11, receiving an undergraduate degree from Tufts at 14, and his PhD from Harvard when he was 18), Wiener was a superstar. He had a wide range of interests and contributed to many fields of study including physics, communication theory, quantum theory and more. During World War II he was invited to work on a project on gunfire control that led to his development of the new field of cybernetics. He not only coined the term but also wrote the definitive, eponymous book about the topic: Cybernetics: or, Control and Communication in the Animal and the Machine.

In many ways, cybernetics was the right word for this conference that McCarthy was organizing. But there were two problems. First, Wiener thought of cybernetics in analog terms not in terms of the digital computer, something McCarthy and the others writing the proposal were passionate about.

Second, and maybe more important, was Wiener’s personality. Wiener was a bit of a character and could be difficult to work with. His “social ineptitude,” as Mary Catherine Bateson writes, could lead to very divergent responses to him as a person: “a complicated mixture of respect for his ideas and exasperation at his lack of a sense of how and when to communicate effectively.” Another colleague writes that his conversation style “was a curious mixture of pomposity and wantonness… He was a poor listener… He spoke many languages but was not easy to understand in any of them.” Yet another had this to say, in an otherwise laudatory obituary, that Wiener would, “snore publicly through a lecture and then ask an awkward question in the discussion” and would often annoy other experts by “proffering information and advice on some field remote from his own.”

All of these facts weighed on McCarthy. He also knew that Wiener was extremely possessive about his ideas, having previously butted heads with Claude Shannon (one of the other organizers of the conference) on their respective contributions in developing information theory. He feared that Wiener might dominate the conference if it were called a cybernetics meeting. As McCarthy wrote (in a document available from his archives at Stanford University):

…  one of the reasons for inventing the term “artificial intelligence” was to escape association with “cybernetics.” Its concentration on analog feedback seemed misguided, and I wished to avoid having either to accept Norbert Wiener as a guru or having to argue with him.

To avoid this and to carve out a distinct identity for the new field, McCarthy decided to coin a new term: artificial intelligence. And a new field was born.

This story is compelling not only for the light it sheds on the personalities shaping the field’s history, but also for its far-reaching consequences. A decision made for seemingly mundane reasons has profoundly influenced not just the discipline itself, but the world at large.

AI today means so many different things, making it, as Joshua Pearson writes in a recent article, “a mystifying umbrella term giving false cohesion to a wide variety of computational techniques.”

Why does it matter?

Unlike “cybernetics,” which was relatively neutral, “intelligence” carries significant philosophical and emotional weight. It implies a direct comparison to human cognition and has led to decades of debate about whether machines can truly “think” or be “intelligent.” This anthropomorphic framing has shaped public expectations, research directions, and ethical discussions in ways that might have been very different had the field retained a more neutral name like cybernetics. The choice of “artificial intelligence” has thus profoundly influenced how we conceptualize, develop, and interact with intelligent machines, demonstrating the enormous power that naming can have in shaping a field’s trajectory and public perception.

Words matter.

A similar argument can be made for the word “hallucination” – the fact that Large Language Models often, extremely confidently, make up stuff out of whole cloth. As Pearson writes, “This wildly evocative term pulls associated concepts of cognition, perception, intentionality, and consciousness into our attempt to understand LLMs and their products, making a murky subject even harder to navigate.” They give the impression as if these errors are accidental and hide the fact that ALL that these LLMs do is make stuff up. They align sometimes, based on the vast nature of their training data, with what we think is right, but at heart they are bullshit machines.

Words matter.

Some possible futures

Coming back to cybernetics v.s. AI what would the world have looked like if McCarthy had gone with cybernetics instead of AI. My guess is the term would have remained a bit niche, and have had less social stickiness and hence virality and I think our perception and approach to the field might have been significantly different.

Here are some possible alternatives that may have emerged (as suggested by Claude.AI):

• Instead of “AI ethics,” we might discuss “cybernetic ethics,” which could focus more on the interplay between human and machine systems rather than on machine consciousness or rights.
• Rather than debating whether AI can be “conscious” or “intelligent,” we might instead explore “cybernetic consciousness,” emphasizing the integration of machine and biological information processing.
• The phrase “artificial general intelligence” (AGI) might be replaced with “comprehensive cybernetic systems,” shifting focus from human-like cognition to advanced, interconnected control systems.
• Instead of “machine learning,” we might use “cybernetic adaptation,” emphasizing the system’s ability to adjust based on feedback rather than implying human-like learning.
• “Neural networks” might be called “cybernetic networks,” potentially leading to different analogies and inspirations for their structure and function.
• The fear of an “AI takeover” might instead be framed as “cybernetic imbalance,” focusing on the disruption of human-machine equilibrium rather than sentient machines rebelling.
• Rather than “AI assistants” like Siri or Alexa, we might have “cybernetic interfaces,” emphasizing their role as mediators between humans and information systems.
• The concept of “artificial emotional intelligence” might be reframed as “cybernetic empathy modeling,” focusing on system responses rather than implying machines have emotions.
• Instead of “autonomous AI,” we might discuss “self-regulating cybernetic systems,” emphasizing control theory rather than independence.
• The field of “AI alignment” might be known as “cybernetic harmony,” focusing on optimizing human-machine interactions rather than aligning machine goals with human values.

These examples demonstrate how the term “cybernetics” could have led to a more systems-oriented, less anthropomorphic approach to the field. It might have resulted in less hype and fear around machine consciousness and more focus on the integration and optimization of human-machine systems.

What would the world have been if we had gone that route? How would we respond to these large language models if we called them Cybernetic models? Would OpenAI be called OpenCybernetics? Would our expectations, fears, and aspirations for technology be different? Would we have developed different kinds of systems or approached their development and integration into society differently? How might this alternative framing have influenced public perception, policy decisions, and the direction of research in this field?

---

Postscript: The goal of the Dartmouth conference was to lay the foundations of this new field. The organizers truly believed that they would achieve human-like intelligence in a few years. Sadly, that was not to be. Intelligence turned out to be a tougher nut to crack. I will let John McCarthy have the last word on this.

“You know,” he said, “intelligence was harder than we thought.”

---

Postscript (July 12, 2024):

The day after I wrote and published this post I came across an article in MIT Technology review titled “What is AI?” This is a great piece for a range of reasons and I do think anybody interested in the current state of AI should read it. More pertinent to this post is that in recounting the history of AI this article goes into greater detail about all the other names that had been discussed before McCarthy decided to go with “artificial intelligence.” Some of the alternatives that were explored were, apart from “cybernetics” were “automata studies,” “complex information processing,” “engineering psychology,” “applied epistemology,” “non-numerical computing,” “neuraldynamics,” “advanced automatic programming,” and “hypothetical automata.” This aspect of this history of the term was not something I was aware of when I wrote my post – and I would have definitely included some of the other possibilities. Each of these names opens some conceptual doors while closing off others, speaking to possible alternative futures of this technology and how we understand it and its relationship to us.

The article also led me to thesis by Jonathan Penn titled: Inventing Intelligence: On the History of Complex Information Processing and Artificial Intelligence in the United States in the Mid-Twentieth Century, in which he adds another layer to the discussion in establishing…

… continuities with, and borrowings from, management science and operations research (Simon), Hayekian economics and instrumentalist statistics (Rosenblatt), automatic coding techniques and pedagogy (McCarthy), and cybernetics (Minsky), along with the broadscale mobilization of Cold War-era civilian-led military science generally.