Artificial intelligence models such as GPT-3 passed the Turing test.
"it already passed the Turing test quite a long time ago like like ChatGPT-3" (said at 0:56:50)
Empirical evaluations and philosophical assessments show that GPT-3 (and its successor GPT-3.5) did not pass the Turing test. In randomized, controlled Turing test experiments, GPT-3.5 was identified as human only ~20% of the time—performing on par with or worse than ELIZA (22%) and well below the human baseline (~66–67%). Earlier formal evaluations of base GPT-3 similarly showed that it routinely failed semantic questioning and tests of human-like intelligence. Only later models, such as GPT-4 and GPT-4o, have achieved pass rates approaching or matching human baselines in two-player Turing tests.
- contradicts: GPT-3: Its Nature, Scope, Limits, and Consequences (Minds and Machines 2020)
"We expand the analysis to present three tests based on mathematical, semantic (that is, the Turing Test), and ethical questions and show that GPT-3 is not designed to pass any of them." (abstract, results, passage verified)
openalexfull study (doi) - contradicts: Does GPT-4 pass the Turing test? (? 2024)
"The best-performing GPT-4 prompt passed in 49.7% of games, outperforming ELIZA (22%) and GPT-3.5 (20%), but falling short of the baseline set by human participants (66%)." (abstract, results, passage verified)
openalexfull study (doi)