How the Test Works
An interrogator has text conversations with two parties: one a human, one a computer. If the interrogator cannot reliably tell which is the machine after a sufficient interrogation, the computer is said to have passed the test. Turing predicted that by the year 2000, a computer would be able to fool 30% of interrogators after 5 minutes of conversation. He was roughly correct — though it took until 2014 for a chatbot to credibly claim to do so.
Criticisms of the Turing Test
Philosopher John Searle's Chinese Room argument (1980) is the most famous critique. Searle imagined a person locked in a room with a rulebook for responding to Chinese symbols, who responds correctly to Chinese questions without understanding Chinese. The room "passes" the Turing test for Chinese but understands nothing. Searle argued this shows that manipulating symbols — what computers do — is not the same as understanding.
Has Any AI Passed the Turing Test?
In 2014, a chatbot named Eugene Goostman — simulating a 13-year-old Ukrainian boy — convinced 33% of judges at a competition that it was human, prompting claims of "passing." Most AI researchers rejected this framing: the chatbot used its fictional persona to excuse ignorance, not demonstrate intelligence. Modern large language models are far more convincing but no formal Turing Test competition consensus exists.
Why It Still Matters
The Turing Test shifted the question of machine intelligence from metaphysics to behavior — from "what is thinking?" to "what can a system do?" It set the agenda for AI research for decades and remains a useful conceptual anchor for debates about machine consciousness, even as most AI researchers have moved to domain-specific benchmarks.
Questions Fréquentes
What is the Turing Test?
The Turing Test (Imitation Game) proposes that if a computer can convince a human interrogator that it is human through text conversation, it can be considered to be 'thinking.' It was proposed by Alan Turing in 1950.
Has any AI passed the Turing Test?
In 2014, the chatbot Eugene Goostman claimed to pass by fooling 33% of judges, but most AI researchers rejected the claim as the chatbot used tricks (pretending to be a non-native speaker child) rather than genuine intelligence. Modern LLMs are far more capable but no definitive test has been formally passed.
What is the Chinese Room argument?
John Searle (1980) argued that a person following rules to manipulate Chinese symbols could answer questions correctly without understanding Chinese — just as computers manipulate symbols without understanding them. The argument challenges whether symbol manipulation can constitute genuine understanding.