Elon Musk discusses new company xAI, ‘superintelligence,’ and China

As a result, he decided to “grow an AI in a good way”. For Musk, this means training him in such a way as to be “curious to the extreme, truth-seeking to the extreme.”

“I think an AI that’s alien to the extreme, that’s just trying to kind of make sense of the universe, would be, I think, pro-human,” Musk said. “From the point of view that humanity is more interesting than not being human.”

Musk has said he prefers this approach to “trying to explicitly program morality into AI.” He said this approach requires judgment as to whose morals are programmed into the system and also creates a “reversal problem” where morals can be turned on their head.

As far as technology regulation goes, Musk has suggested that policymakers spend a few years getting to know the technology before moving on to “oversight, in consultation with industry.” He also suggested that some sort of industry group or self-regulatory body similar to the Motion Picture Association would be useful.

The three spent much of the discussion focused on the threat of China’s supremacy over the United States in artificial intelligence and the possibility of some kind of military confrontation over the sovereignty of Taiwan, an autonomous region that the Chinese government considers part of its territory. province.

Musk said that in his conversations with the top leadership in China on his recent trip there, he spent a great deal of time discussing AI security. He said the idea that “digital superintelligence” could replace the Chinese Communist Party itself seemed to be resonating.

“No government wants to find itself defeated by digital superintelligence,” he said. “So I think they’re already taking action on the regulatory front and they’re concerned about this as a risk.”

Musk even said he believes the Chinese government would be open to cooperation in an international framework on AI regulation.

Gallagher, who chairs the Select Committee on China, backtracked on the notion that the CPC could be a constructive member of such an international framework. He warned that even if they take Musk’s warnings seriously, he fears that it will not be enough to slow their AI efforts and that Chinese leader Xi Jinping will use them to promote “totalitarian control.”

After changing the subject briefly, Musk reverted to Gallagher’s thoughts and described himself as “pro-Chinese.” Musk has admitted that he has “some vested interests in China” but ultimately believes that “China is underappreciated” and that “the people of China are really amazing.” Gallagher later said that he “fully supports the Chinese people”, but that it is the ruling party with which he disagrees.

While tensions between the US and China remain high, the two powers have made some diplomatic progress by resuming personal contacts, with Treasury Secretary Janet Yellen recently taking a trip there and US climate envoy John Kerry preparing to do so later this month.

However, Musk expected that resolving the Taiwan issue would be difficult.

“I already have this theory about prediction, that the most interesting outcome — as seen by a third party, not the participants — is the most likely,” Musk said. “Suggesting that it’s probably warming in the Pacific Ocean. Hopefully it’s not too hot. But it’s going to get hot.”

Musk expected the world to see digital superintelligence, or a form of artificial intelligence “smarter than any human being at anything” within five or six years.

“This is not necessarily smarter than the sum of all human beings,” Musk said. “That’s a higher standard for being smarter than many humans. Especially considering it’s the sum of all machine-augmented humans we’ll have computers, phones, and software applications. Computers are not yet integrated with us.”

“If you set odds, I think it’s more likely to be a positive scenario than a bad scenario,” he said. “It’s just that a bad scenario is not 0% and we want to do everything we can to reduce the likelihood of a bad outcome with AI.”

Musk at one point described the future as being like a TV series.

“If this were a Netflix series or something,” he said, “I would say the season finale would be a confrontation between the West and China, and the series finale would be AGI,” or Artificial General Intelligence.