But in Herzog’s documentary, Musk’s fear isn’t all powerful, Terminator-style A.I. robots. Instead, the tech mogul is worried about A.I. that does whatever it takes to accomplish its task.
“The biggest risk is not that A.I. will develop a will of its own,” Musk says in a short clip of the new film obtained by Fortune. “But rather it will follow the will of its utility function or optimization function.”
Herzog, the man who made Grizzly Man, is more attuned to nature than future technologies. Judging from early trailers and reviews from the documentary’s premiere at the Sundance Movie Festival in January, the whole documentary has a bit of a Luddite tilt.
It makes sense that Herzog would reach out to Musk for a deeper understanding of what the world will be like in the future. Musk is, after all, the man who is shaping the world’s autonomous electric vehicles, space travel, and Reddit-learning A.I. In the case of A.I., Musk sees a threat around every line of code.
“If it is not well thought out, even if its intent is benign,” Musk says, “it could have quite a bad outcome.”
Take A.I. that manages a hedge fund or private equity fund, Musk suggests. The A.I.’s sole job is to maximize the fund’s portfolio. Sounds simple and innocent enough. But, say, the A.I. realizes the best way to make a Michael Burry/The Big Short amount of money is to short consumer stocks, go long on defense stocks, and then figure out a way to start a war (presumably through its A.I. friends in the Department of Defense)?
It’s not that out there of an idea. It’s a bit of a slippery slope, but not unimaginable. Of course, it won’t happen if humans just take Musk’s advice and use neural lace to seamlessly integrate with A.I.
Lo and Behold will be in select theaters and available on Amazon, Google Play, and iTunes on August 19.