OpenAI’s GPT-4, a generative artificial intelligence (AI) model, has passed the Turing Test, according to Ethereum co-founder Vitalik Buterin.
Turing Test (Turing Test) is a test of the ability of artificial intelligence to see if it can do well close to that of humans or not. The term was coined after famous mathematician Alan Turing proposed the test in the 1950s.
The test method is to have the referee talk to people and artificial intelligence. (will be speaking or printing It doesn’t focus on communication abilities) and decides whether the conversation is with a person or an artificial intelligence. And if it cannot be differentiated, it means that the artificial intelligence has passed the Turing Test.
According to Turing, at that time, AI systems that could create messages that trick humans into thinking they were conversing with another human would demonstrate the ability to “think.”
Nearly 75 years later, Vitalik Buteri has interpretedresearchA recent preprint of the University of California, San Diego. It indicates that artificial intelligence has finally passed the Turing test.
Researchers from the University of California San Diego recently published a pre-print report titled “People were unable to distinguish GPT-4 from humans in the Turing test,” and they had approximately 500 human testers interact with humans and AI models in a blind test to verify that They can determine which is a person or an artificial intelligence.
Research found that humans mistook GPT-4 for “human” 56% of the time, which means that machines tricked humans into thinking they were human.
According to Buterin, an AI system that can fool more than half of humans qualifies as passing the Turing Test.
Buterin validated his statement by saying: “Not quite yet. Because humans were assumed to be 66% human vs. 54% for bots, the 12% difference is very small.”
refer : cointelegraph.com
picture cryptoslate.com
The post Vitalik Buterin says OpenAI’s GPT-4 passes the Turing Test appeared first on Bitcoin Addict.