A group of researchers claim that GPT-4, the latest version of the ChatGPT language model, has passed the Turing Test, fooling people into thinking they were interacting with a human. The Turing Test evaluates whether an artificial intelligence can impersonate a human and deceive others, as proposed by Alan Turing in 1950. The study conducted by the University of San Diego involved testing GPT-4 and GPT-3.5, with GPT-4 successfully fooling participants half the time. Despite this, humans still outperformed in identifying other humans in the test. Language models like GPT-4 are trained to understand, create, and interact with text, based on neural networks and Transformers. OpenAI recently released GPT-4, an optimized version of ChatGPT, capable of generating longer and more cohesive texts, as well as providing up-to-date information from the internet. The study’s findings suggest that GPT-4 has advanced to the point of passing the Turing Test, marking a significant milestone in artificial intelligence development.
Source link
Source link: https://www.yourweather.co.uk/news/trending/a-new-era-for-chatgpt-gpt-4-may-have-passed-the-famous-turing-test.html
Is ChatGPT entering a new era with GPT-4? #AIRevolution
