Elon Musk, a billionaire businessman, announced on Monday that he would launch an artificial-intelligence (AI) platform he has named “TruthGPT”, to compete with the offerings of Microsoft and Google.
He accused Microsoft-backed OpenAI – the company behind the chatbot sensation ChatGPT – of “training AI to lie”. OpenAI, he said, has become a closed-source, “for-profit”, organisation that is “closely aligned with Microsoft”.
He also accused Larry Page (co-founder of Google) of not taking AI security seriously.
Musk stated in a Monday interview with Tucker Carlson of Fox News Channel that “I am going to start a project which I will call TruthGPT, or an AI maximum truth-seeking system that tries understanding the nature and the origins of the universe.”
He said TruthGPT was “possibly the best way to safety”, and that it would “unlikely annihilate human beings”.
It’s just that it started late. Musk said, “I will create a third alternative.”
Musk, OpenAI and Page didn’t immediately respond to Reuters requests for comment.
Reuters reported that Musk is poaching AI experts from Alphabet Inc.’s. Google in order to launch a rival startup, OpenAI.
According to a filing with the Nevada state government, Musk registered a Nevada-based company named X.AI Corp last month. Musk was listed as the only director, and Jared Birchall as secretary. Birchall is the managing director of Musk’s family office.
Musk, along with a group artificial intelligence experts, and executives from the industry called for to take a 6-month break before developing systems that are more powerful than OpenAI’s recently launched GPT-4. They cited potential risks to society.
Musk also repeated his AI warnings during the interview Carlson. According to the excerpts, “AI is more harmful than, for example, mismanaged airplane design, production maintenance, or bad car manufacturing.”
He said that it could lead to the destruction of civilisation.
He claimed, for instance, that an AI super-intelligent could write extremely well and possibly manipulate public opinion.
He tweeted on the weekend that, when Barack Obama was President of the United States, he met with him and told him Washington should “encourage AI regulations”.
Musk founded OpenAI in 2015. He resigned from the board of directors in 2018. Musk tweeted in 2019 that he was leaving OpenAI to concentrate on Tesla and SpaceX.
He tweeted that he left OpenAI because “Tesla competed for the same people that OpenAI wanted & I did not agree with what OpenAI wanted to do.”
Musk, the CEO of Tesla, SpaceX and now Twitter, is also CEO. He bought Twitter for $44 billion in 2013.
Musk told Fox News that he valued Twitter “less than half” the price of its acquisition in a recent interview.
Microsoft Corp announced in January a multi-billion-dollar investment in OpenAI. This intensified the competition with Google and fueled the race for to attract AI funding.
Musk signed a letter last month with more than 500 other experts, calling for an AI moratorium for six months.
The letter said that “Advanced Artificial Intelligence could be a major change in the history and development of life on Earth and it should be managed and planned with the same care and resources.” This level of planning and managing is unfortunately not occurring, despite the fact that AI labs have been locked in a race out-of-control to develop and deploy digital minds which no one, including their creators, can understand, predict or reliably control.