Concept · Chapter 1: What Is Artificial Intelligence?
The Turing Test
The Turing test replaces the vague question 'Can machines think?' with a concrete one: can a machine's conversation be told apart from a human's?
The problem
'Thinking' and 'intelligence' have no agreed definitions, so arguments about machine intelligence went in circles.
The solution
Judge by behaviour: if an interrogator, conversing by text, can't reliably distinguish the machine from a person, treat it as intelligent for practical purposes.
The consequence
It gave AI an operational goal and a durable debate — and today's fluent language models show how much, and how little, passing for human really tells you.
The idea
In 1950 Alan Turing sidestepped a philosophical swamp. Instead of defining "thinking", he described a game: an interrogator exchanges typed messages with two hidden players, one human and one machine, and tries to tell which is which. If the machine can't be reliably picked out, Turing argued, we have no better grounds for denying it intelligence than we have for granting it to other people.
Why it still matters
- It defines intelligence by behaviour. That operational stance — judge the system by what it does — runs through all of AI, including today's benchmarks.
- It is easy to game. Weizenbaum's 1966 ELIZA used simple pattern-matching, yet some users treated it as if it understood them Established. Sounding human and being capable are different things.
- The debate is live again. Modern chatbots can hold long, fluent conversations. Whether that fluency reflects understanding is contested, and most researchers now prefer many specific capability tests over one conversational one Interpretation.
What to remember
- Turing (1950) proposed the imitation game as a replacement for 'Can machines think?'
- It tests behaviour (conversation), not internal mechanism.
- Critics (e.g. Searle's 'Chinese Room', 1980) argue behaviour isn't understanding.
- Fluent text is easy to mistake for intelligence — ELIZA showed this in 1966.
Key papers
Computing Machinery and Intelligence
A. M. Turing · 1950 · Mind
Replaced the unanswerable question 'Can machines think?' with a testable one — the imitation game — and anticipated learning machines decades early.
How to read it: Very readable, even witty. Section 7, 'Learning Machines', is remarkably prescient — read it after Chapter 1.