Tesla and Twitter owner Elon Musk makes the announcement in an interview with Fox News.

At the same time, he criticizes Open AI, the company backed by Microsoft and behind ChatGPT. According to Musk, the software for the high-profile AI service is programmed by "left-wing experts" who "taught the chatbot to lie."

"Truth-seeking AI"

On top of that, he accuses Google CEO Larry Page of not taking AI security seriously.

"I'm going to start something I call TruthGPT, or a maximally truth-seeking AI that tries to understand the nature of the universe," he says.

"I think this is perhaps the best path to security in the sense that an AI that cares about understanding the universe is unlikely to annihilate humanity because we are an interesting part of the universe.

Sources told Reuters that he has recruited AI researchers from Google to the project.

In the past, Musk has warned of the consequences of AI and called for a six-month pause in the development of new advanced systems until society has "caught up". In the interview, he reiterates his warnings, claiming that AI has "the potential, however small you think the probability is, to destroy civilization."

Co-founder of Open AI

Musk was one of the co-founders of Open AI when the company was created in 2015. However, he left the board of directors in 2018.

"I had to focus on solving a painful amount of technology and manufacturing concerns at Tesla, most notably, and Space X," he wrote in 2019 on Twitter about the reasons behind the decision.

"Also, Tesla was competing for some of the same people as Open AI and I disagreed with some of the things that Open AI wanted to do. Add it all together and it was simply better to leave on good terms."

Javascript is disabled

Javascript must be turned on to play video

Learn more about browser support

From the boys' room in South Africa to rocket launches into Mars, this is Elon Musk. Photo: Patrick Pleul/Pool/AP/TT