John McCarthy and the myth of “AI”: why CIP is a more accurate term for the technology

The debate about whether AI will replace humans has largely grown out of the name of the technology itself. If we remove the flashy label "artificial intelligence," we're left with a more mundane description: systems of complex information processing.

How the Term Emerged

Returning to the 1950s helps us understand where today's "human versus machine" conflict comes from. The foundation for the future discipline was laid by Alan Turing and Claude Shannon, but the term artificial intelligence itself was proposed by John McCarthy in 1956. He organized the Dartmouth Summer Research Project on Artificial Intelligence, which later became the founding moment of the entire field.

In this history, what matters is not just the scientific content but the packaging of the idea. McCarthy needed a term that would distinguish the new direction from Norbert Wiener's cybernetics, sound ambitious, and help attract attention. Essentially, it was a powerful research brand: the name promised not just computation, but almost the creation of mind.

The ambition matched. McCarthy believed that noticeable progress toward human-level capability could be achieved very quickly, almost within a single research summer. But from the start, it became clear that gathering scientists under one umbrella was simpler on paper than in reality: the invited specialists pursued their own research, and overall coordination was difficult.

Naming and Expectations

Such word choice influenced the perception of technologies for decades to come. When a program is called intelligence, it is automatically ascribed human properties: intention, understanding, creativity, independent will. Hence the constant question of whether it will "replace" a specialist, as though we're talking about a new labor market participant rather than a set of methods.

"Plant the flag on the mast" — this is how you could describe the choice of a resonant name and the struggle for attention to the new field.

This decision also had a practical side. Under such a striking term, it was easier to secure funding: McCarthy attracted a $7,500 grant from the Rockefeller Foundation. For the mid-1950s, this was an important resource, but along with the money came inflated expectations. If a project is called "artificial intelligence," the public and investors begin to expect not tools, but a digital rival to humans.

Why Newell and Simon Disagreed

Against this background, the position of Allen Newell and Herbert Simon is particularly interesting. Both stood at the origins of the field, later received the highest scientific honors, and both opposed the new discipline being built around the term "AI." Instead, they used a more precise definition—complex information processing.

This shift seems purely semantic, but in fact it changes the frame of the conversation. If a system is engaged in complex information processing, it is easier to evaluate it by specific operations rather than by fantasies about "machine minds." Then the discussion is not about competition with humans in general, but about which specific functions the technology can speed up, cheapen, or automate.

• finding patterns in large datasets
• comparing options and ranking solutions
• quickly extracting facts from documents
• generating drafts, forecasts, and hypotheses
• supporting specialists in routine analytical tasks

This does not eliminate the risk of displacement, but it makes it more concrete. It is not "people in general" that are at risk, but repeatable functions within professions. Where work boils down to templated information sorting, automation can indeed displace the worker. Where context, responsibility, task setting, and result verification are needed, the human role remains central.

What This Means

The term "artificial intelligence" remains convenient for the market, media, and investors, but it distorts expectations. The formulation "complex information processing" sounds less striking, but more accurately describes what modern systems actually do. For a sought-after specialist, this is bad news only where their work is entirely routine; in all other cases, we're talking more about a redistribution of roles than a direct replacement of humans.