Tesla and SpaceX owner Elon Musk has announced plans to create a “TruthGPT” chatbot that he claims will counteract political bias in other chatbots. In an interview with Fox News, he accused OpenAI’s ChatGPT of being “trained to be politically correct” and said that his new chatbot would be a “maximum truth-seeking AI that tries to understand the nature of the universe”. Musk’s announcement follows his call for a six-month pause in the development of powerful AI systems last month, due to their potential risks to society. He has also registered a new company, X.AI Corp, in Nevada.

Musk stated that he believes AI is “more dangerous” than the technology behind SpaceX’s rockets or Tesla’s cars and called for regulation of the sector. He said that governments should take action and praised the Chinese government’s draft rules on generative AI services, which are expected to be heavily restricted. Italy has already banned ChatGPT and Europol, the EU law enforcement agency, has warned that the chatbot may be used by criminals to spread disinformation.

OpenAI’s ChatGPT has been hugely successful, with over 100 million monthly users. Companies including Microsoft and Expedia have also used the chatbot’s technology in their own products. However, there have been concerns over the accuracy of information provided by such models, as they have been shown to occasionally make false claims. Musk said that his new chatbot would seek to understand the “nature of the universe” and may be less biased than ChatGPT.

Musk has a history of involvement with OpenAI, co-founding the company in 2015 and stepping down from the board in 2018. His criticism of OpenAI has led him to poach AI researchers from Google to work on his new startup. The creation of X.AI Corp has also been seen as part of this process. Despite his concerns about AI’s dangers, Musk said that he believes it is “unlikely to annihilate humans” and that TruthGPT “might be the best path to safety”.