Scientists at UCSD have revealed that leading AI chatbots, GPT-4.5 and LLaMa, have successfully passed the Turing test, showing significant advancements in AI capabilities. The Turing test, developed by Alan Turing, assesses whether a machine's responses can be distinguished from those of a human. In research involving 126 students and other participants engaging in conversations with human and AI counterparts, GPT-4.5 was identified incorrectly as human 73% of the time, exceeding the correct identification of real humans, which signifies a notable milestone in AI development.
The results constitute the first empirical evidence that any artificial system passes a standard three-party Turing test, demonstrating AI's ability to imitate human conversation.
Participants had five-minute online conversations with another human and one of the AIs and had to judge which they thought was human, highlighting AI's sophistication.
Collection
[
|
...
]