The billionaire founder of Tesla Motors and SpaceX, Elon Musk, has said before that a hostile artificial intelligence (AI) is a real concern of his.
He even called it probably “our biggest existential threat,” talking at an MIT symposium two years ago.
“With artificial intelligence, we are summoning the demon,” Musk said, without going into much detail about why did he thinks this way.
On June 1, Musk elaborated while speaking at The Verge’s Kara Swisher and Walt Mossberg at the Recode’s Code Conference 2016.
So what’s Musk’s problem with AI?
First, he indicated that AI is on its way to becoming massively more intelligent than humans—not just equivalent or more intelligent.
“If you assume any rate of advancement in AI, we will be left behind by a lot,” he said.
The first order of business is, “if we create some digital super-intelligence that exceeds us in every way by a lot, it’s very important that that’d be benign.”
Two years ago, Musk mentioned regulation and oversight, but his approach has changed since then.
“I think it’s important if we have this incredible power of AI that it not be concentrated in the hands of a few and potentially lead to a world that we don’t want,” he said.
“If instead AI power is broadly distributed and to degree that we can link AI power to each individual’s will,” meaning everybody would have control over their own powerful AI servant, “then if somebody did try to do something really terrible then the collective will of others could overcome that bad actor.”
For that reason, Musk was one of the initiators of the OpenAI last December. The non-profit received pledges for $1 billion to develop AI and make it available for free.
That, of course, doesn’t mean somebody wouldn’t develop much more powerful AI privately.
Musk acknowledged that possibility, but he said he’s only worried about one company doing so.
“I won’t name a name,” he said. “But there is only one.”
Based on the context, The Verge reported Musk probably meant Google, who bought British AI company DeepMind in 2014 and continues to make significant progress with the technology.
The question would then be if the OpenAI or any freely distributed AI will be able to match the power of the ones developed privately.
Yet that still doesn’t solve the first issue Musk raised—the progress of AI.
“If you have ultra-intelligent AI, we would be so far below them in intelligence that we would be like a pet,” Musk said. “We’ll be like the house cat.”
“I mean, I don’t love the idea of being a house cat,” he said.