Speaking at SXSW, SpaceX and Tesla founder Elon Musk warned that the threat posed by unregulated artificial intelligence is bigger than that of nuclear weapons.
“Mark my words,” Musk said, “AI is much more dangerous than nukes. So why do we have no regulatory oversight?”
He believes that a public regulatory body has to be created that would need “insight and oversight” to ensure that everybody working with AI does so in a way that is “symbiotic with humanity.” If such a body is not created, he said, AI gone wild could cause a third world war… and the end of mankind:
However, even coming up with the safety parameters would , Musk said. If the utility function of artificial intelligence is to maximise happiness of humans, a super-intelligent AI might decide that the best way to do that is to capture all humans and inject their brains with dopamine and serotonin.
Musk proposed that digital intelligence should instead be directed to maximise “the freedom of action of humanity.”
Although Musk certainly has skin in the game — Musk frequently uses “a third world war” as the number one reason for wanting to colonize Mars with SpaceX — that doesn’t negate the importance of heeding his warning. More than most people, Musk knows what his fellow innovators are working on. If he is sincerely afraid that AI will take a turn for the worse and pose a serious threat to humanity, we’d better pay attention.