Papers / Computing Machinery and Intelligence / Section 03 · The Core Idea
Section 03

The Core Idea

Foundation of Artificial Intelligence Computing Machinery and Intelligence 1950

The Core Idea

The Imitation Game

Turing’s key insight was to replace a philosophical debate with a practical game.

He described a game with three players:

  • Player A — a man
  • Player B — a woman
  • Player C — an interrogator, who cannot see or hear A or B and can only communicate with them by typed messages

The interrogator’s job is to figure out which one is the man and which is the woman. Player A is trying to fool the interrogator (claiming to be a woman). Player B is trying to help the interrogator (telling the truth about being a woman).

Turing then asks: what happens if we replace Player A (the man) with a machine? Can the machine fool the interrogator as effectively as the man could?

This is the Imitation Game — and it is what we now call the Turing Test.

The simplicity of this idea is genius. Instead of arguing about consciousness or souls or the nature of the mind, Turing says: forget all that. Let the machine compete. If it can hold up its end of a conversation — if an intelligent human cannot reliably tell the difference between talking to the machine and talking to a person — then for all practical purposes, the machine is thinking.

The analogy: the impersonation competition

Here is an Indian analogy that captures the idea.

Imagine a reality show called Voice of India. The format is unusual: contestants are hidden behind a screen. The audience can only hear them speak. Each contestant’s job is to impersonate a famous person — a film star, a politician, a cricket commentator.

The judges must listen and decide: is this the real person, or an impersonator?

Now imagine that instead of a human impersonator, an AI voice is competing. The AI has been trained on thousands of hours of that famous person’s voice, speech patterns, and vocabulary.

If the AI can fool the judges just as often as the human impersonators can — if there is no reliable way to tell the AI apart from a person — then Turing would say: for all practical purposes, the AI can imitate the person. Whether there is something it “feels like” from the inside to be the AI is a separate question, and perhaps an unanswerable one.

A second analogy: the exam without a face

Think of a competitive entrance exam — say, JEE Mains. All the answers are written. You never know if the person answering is young or old, from a village or a city, nervous or confident. The examiner only sees the answer sheet.

If a machine could write an answer sheet indistinguishable from that of a brilliant human student, would we say the machine “understood” the subject?

Turing would say: that is exactly the right question. And we should stop arguing about what is “really” happening inside the machine and judge it by the answer sheet.

How is this different from what came before?

Before Turing, the question “can machines think?” was treated as a philosophical question — answered by introspection, theology, or intuition. Turing transformed it into an empirical question — one that can be investigated by experiment and observation.

This is a massive shift. Instead of philosophers debating in armchairs, you now have engineers with something to build toward. A target. A test. A definition of success.

The field of artificial intelligence — which was officially named in 1956 at the Dartmouth Conference, six years after this paper — was born partly from this shift. If you have a test, you can try to pass the test. And trying to pass the test means building things, running experiments, measuring results.

That is how science works.

The nine objections Turing answered

A remarkable part of the paper is that Turing did not just propose the test. He then systematically went through nine objections that intelligent people might raise, and answered them one by one. This is worth noting because it shows the depth of his thinking.

The nine objections were:

  1. The Theological Objection — Only God can give a soul to a thinking being; machines have no souls; therefore machines cannot think.
  2. The “Heads in the Sand” Objection — The consequences of machines thinking would be too terrible; therefore we should not consider it possible.
  3. The Mathematical Objection — Gödel’s theorem shows that any formal mathematical system has limits; therefore machines have limits that minds do not.
  4. The Argument from Consciousness — A machine cannot truly think because it cannot feel — it has no genuine experience.
  5. The Argument from Various Disabilities — Machines cannot be kind, creative, romantic, funny, grieving — they cannot “really” do any of these things.
  6. Lady Lovelace’s Objection — Ada Lovelace (who worked with Charles Babbage) said that a machine can only do what it is instructed; it cannot surprise us.
  7. The Argument from Continuity in the Nervous System — The nervous system is a continuous analogue system; digital computers are discrete; they are fundamentally different.
  8. The Argument from Informality of Behaviour — Human behaviour cannot be captured by rules; machines can only follow rules; therefore machines cannot replicate human behaviour.
  9. The Argument from Extra-Sensory Perception — Telepathy, clairvoyance, etc. exist; machines cannot have these; therefore machines and humans are different.

Turing’s responses to most of these objections remain among the sharpest thinking in the entire history of AI philosophy. The debate he started in 1950 is still going, with the same objections and refinements of the same responses.

Next: How It Works →

← Section 02: The Problem