GPT-4.5 and Meta's AI Tested in Turing Test

Recent studies conducted by researchers at the University of California, San Diego, reveal that OpenAI's GPT-4.5 has successfully passed the Turing Test, convincing human participants it was human 73% of the time, surpassing the performance of actual humans. The study also evaluated Meta's LLaMa-3.1, which was less convincing but still managed to deceive participants significantly. This success has sparked renewed debate about the validity of the Turing Test as a measure of artificial intelligence, with some experts arguing it tests human assumptions rather than true machine intelligence. The results demonstrate the advanced capabilities of current AI models in mimicking human conversation, raising questions about our ability to discern AI from human interactions.