Elon Musk is on the mic warning us about AI once more. This is the inventor (“genius, billionaire, playboy, philanthropist…”) who wants to put a dude on Mars in five years, and had no problems specifying the space suits for SpaceX’s Mars missions had to look “badass.” This is the guy who conceived and developed an electric car with such “cartoonishly fast” acceleration that it blew auto reviewers’ hair back worldwide. He has five kids and keeps them in his life: a daunting task even when you aren’t an international figure.
So why is Musk so worried? What does it take to rattle an otherwise fearless man? With tongue held firmly in cheek, the answer seems to be “something smarter than he is.”
“I have exposure to the very cutting edge AI, and I think people should be really concerned about it,” Musk said at a recent meeting of the National Governors Association. “I keep sounding the alarm bell, but until people see robots going down the street killing people, they don’t know how to react, because it seems so ethereal.
“AI is the rare case where I think we need to be proactive in regulation instead of reactive,” Musk added. “Because I think by the time we are reactive in AI regulation, it’ll be too late. AI is a fundamental risk to the existence of human civilization.”
Mr. Musk spoke for about 45 minutes at the conference, and if you like, you can watch his full remarks below:
This is by no means Musk’s first public warning. But it’s not like he’s an evil genius about to loose a malevolent AI upon an unsuspecting populace. (Five kids. Not enough time.) So if he doesn’t have the killer AI, and we don’t have the killer AI, and North Korea seems not to have the killer AI…what’s the big worry? Where are the murderbots? It doesn’t much look like the sky is falling.
But we humans are good at weaponizing things. We don’t need no stinkin’ killbots. No, far from rampaging murder machines, it might be fake news that brings us down.
Arguably the greatest threat is mass population control via message targeting and propaganda bot armies. ML not a requirement though
— François Chollet (@fchollet) July 16, 2017
Killbots, as it turns out, aren’t what Elon Musk and the other folks worried about AI are actually concerned about. They’re concerned about an intelligence that surpasses ours, with values different than ours.
This is a little like the classic tension between the two fun-filled and family-friendly dystopias of Orwell and Huxley, or maybe a little of both. Killbots are more like an Orwellian endgame, where the technology is the method of nonconsensual control and destruction: a Panopticon via telescreens, or high-tech jails full of terrors. Propaganda bots driving us to mutually assured destruction, though, might be more up Huxley’s alley, with the idea that a gadgetized human herd wreaks its poetic fate by its own assent.
Wouldn’t it be funny if the AI never tried to take control at all? Instead of some sophisticated coup enabled by machine learning and the high-level use of mass surveillance, what if the thing that actually brought us down was the equivalent of AI jerking on the societal steering wheel the instant it got access?
Now read: Scientists develop official guidance on robot ethics