Subject areaArtificial intelligence
Also known asimitation game
The Turing test is the trial proposed by Alan Turing in 1950, in the paper Computing Machinery and Intelligence, to tackle the question "can machines think?".
Turing replaces it with a question that can be checked: a machine passes if, in a written conversation, the interrogator cannot tell it apart from a person.
Turing calls it the imitation game and builds it from three parts:
- the interrogator, who asks questions in writing and sees only the answers;
- the person, who answers honestly;
- the machine, which answers trying to be taken for a person.
The test measures behaviour, meaning what the machine writes, and leaves aside what happens inside it. For this reason it remains the starting point of every discussion on artificial intelligence: a program able to hold a convincing conversation passes the test even when it gets the facts wrong, so in a company a fluent conversation is assessed separately from the accuracy of the answers.
Source: A. M. Turing, Computing Machinery and Intelligence, Mind, 1950.
An example
The case below is constructed, because the term does not yet appear in the articles. A bank tests an automated customer service assistant: a group of employees types questions without knowing whether the program or a colleague is answering. If nobody can tell them apart, the assistant passes a reduced version of the Turing test, and it remains to analyse whether the answers are also correct.
Enrico Giubertoni. Licensed under CC BY-NC 4.0: you may quote, translate and adapt this entry with the author’s name and a link, stating any changes. For commercial use, write to info@enricogiubertoni.com
Spotted an inaccuracy or something that needs updating? Write to info@enricogiubertoni.com and help make this glossary a better place.
What are your business challenges?
Tell us about your priorities and the objectives you want to reach.
You will receive a free, targeted answer within one working day.