Tesla CEO Elon Musk weighed in on the dangers of artificial intelligence (again) today, with a dire warning.
Speaking at the Wall Street Journal’s CEO Summit, Musk said, “There is a risk that advanced AI either eliminates or constrains humanity's growth.” When asked to elaborate on that comment later in the interview, Musk said he was concerned about an “un-benign” scenario related to the hyper-advancement of AI.
“I don’t think AI will try to destroy all of humanity, but it might put us under strict control,” Musk said, though he surmised there is a “non-zero chance of [AI] going full ‘Terminator’,” referencing the deadly time-traveling robots in James Cameron’s 1984 sci-fi film.
More likely, Musk mused, AI could take control of all of forms of technology — computing power, weapons, etc. — for “the safety of all humans,” effectively acting as an “uber-nanny.”
A ‘third horse in the race'
Though Musk is wary about the prospects of AI, that does not mean he is staying out of the business —far from it.
Musk said Microsoft, with its partnership with OpenAI, and Google, with its Bard AI technology, need more competition.
“There should be a significant third horse in the race,” Musk said, noting that he would have “more on that soon.”
Musk went on to say he's is trying to emulate Microsoft’s relationship with OpenAI for his own AI startup, X.AI, by partnering with his other companies, notably Twitter and Tesla.
Beyond partnerships, Musk said a significant amount of cash will be needed to win the AI race, based on the computing power alone. “Minimum is $250 million in server hardware,” he said. Beyond capital, he said three other requirements are talent, compute power, and data. Whichever company has all three will win, he said.
Despite his dire warnings about the advancement of AI, Musk believes Tesla could be one of the big winners in the space, touting the automaker’s huge commitment to AI and machine learning. Musk has said Tesla’s Full-Self Driving software, now in beta, would be end-to-end controlled by a neural network AI model.
Regulation and an 'era of abundance'
Musk said society still had time to control the impact of AI through - believe it or not - regulation.
Musk, one of the more notable signatories of a recent open letter urging for a pause in development of the most advanced AI systems due to AI’s “profound risks to society and humanity," believes now is the time to act.
At the CEO event, Musk, a noted skeptic of government interference, called for regulation in the form of an “insight committee,” which would be comprised of experts and government officials that would meet from time to time and keep tabs on what’s going on in the AI arena.
The insight committee would have the ability to propose rule-making but not make rules, he said, along the lines of the Federal Aviation Administration and Food and Drug Administration, which can only issue rules unilaterally under certain circumstances.
But, Musk predicted, AI won't lead to a completely dystopian future. Instead, it will likely usher in an "era of abundance" by being able to create goods and services very inexpensively. It’s only when the technology is called on for more advanced tasks, Musk surmised, that humanity might be at risk.