OpenAI trained ChatGPT to lie: Elon Musk

CEO Twitter and SpaceX Elon Musk criticised Microsoft-backed artificial intelligence (AI) startup OpenAI — creator of ChatGPT — and accused that they trained the AI chatbot to "lie".

Speaking during an interview with Fox News aired on Monday, he announced that he would be launching his AI platform that he called "TruthGPT" to "challenge the offerings from Microsoft and Google".

While accusing OpenAI of "training the AI to lie", he said: "OpenAI has now become a closed source, 'for-profit' organisation closely allied with Microsoft".

During the interview, he also alleged the co-founder of the Google Larry page for not taking AI safety seriously.

"I'm going to start something which I call 'TruthGPT', or a maximum truth-seeking AI that tries to understand the nature of the universe," Elon the tech-billionaire said during the interview.

TruthGPT might be the best path to safety that would be unlikely to annihilate humans, noted Musk who is also the CEO of Tesla.

He added: "It's simply starting late. But I will try to create a third option."

The Twitter CEO has been looking for AI researchers from Alphabet Google to initiate an OpenAI rival project, reported Reuters citing sources.

In March, he registered a company in Nevada named X.AI Corp. that listed him as the sole director and Jared Birchall as a secretary — the managing director of Musk's family office.

'AI risks humanity'

The development is followed by an open letter written by technology executives, and AI researchers including Elon Musk calling for a pause of six months in building a system that could be far more powerful than OpenAI's ChatGPT 4.

In the letter, it was mentioned that AI labs are currently locked in an "out-of-control race" to develop and deploy machine learning systems "that no one — not even their creators — can understand, predict, or reliably control."

"AI systems with human-competitive intelligence can pose profound risks to society and humanity," said the letter.

"Powerful AI systems should be developed only once we are confident that their effects will be positive and their risks will be manageable", read the letter in which Musk was also among the people asking to halt the speedy development the AI technology.

Elon Musk also cautioned about human-like technology saying: "AI is more dangerous than, say, mismanaged aircraft design or production maintenance or bad car production according to the excerpts."

"It has the potential of civilizational destruction," Musk added.

He went on to say that "For example, a super-intelligent AI can write incredibly well and potentially manipulate public opinions."

In a tweet, over the weekend he said that he had met with former US President Barack Obama when he was president and told him that Washington needed to encourage AI regulation.

In 2015, Musk co-founded OpenAI but stepped down from the board in 2018.

Explaining the reasons in a Tweet he had said: "Tesla was competing for some of the same people as OpenAI [and] I didn't agree with some of what OpenAI team wanted to do."