Aug 10, 2015

Will super-intelligent AI be religious? Some say yes.

In the world of big-budget, blockbuster comic-book movies, it is plausible: a religiously motivated robot.

As imagined by Joss Whedon, the super-smart artificial intelligence at the center of the most recent Avengers movie turns to villainy because of a religious question. Ultron asks about the meaning and purpose of existence, then, frustrated, committed itself to destruction.

"It's our new Frankenstein myth," Whedon said, doing promos for Avengers: Age of Ultron. "We create something in our own image and the thing turns on us. It has that pain of 'Well, why was I made? I want to kill Daddy.'"

Whedon is an atheist whose many projects have frequently explored religious questions. The most recent Avengers villain raises an unusual one: When humans eventually create super-intelligent artificial life, will it be religious?

Some people think it's plausible even in the real world.

Lincoln Cannon, president of the Mormon Transhumanist Association, says it's definitely possible.

In a recent interview, Cannon said:
Of course there are some naive voices among the anti-religious that would like to imagine a technical incompatibility between machine intelligence and religious beliefs, but humans are already proof of concept. I do think we can identify some limits to the possibility space of intelligence in general, based on logic and physics, but religiosity remains clearly within the possibility space.
Religious super-thinking robots may not be a good thing, though. Cannon goes on:
Religious superintelligence may be either the best or the worst kind of super intelligence -- sublimely compassionate or horribly oppressive. I like to think of religion as applied esthetics, the most powerful social technology for provoking strenuous action toward a common goal. As such, it's not inherently good or evil. It's just power, to be used for good or evil, as it clearly has been used for both historically.

While the particular forms of religion will continue to evolve, the general function of religion seems unlikely to go away. So, as we do with all powerful technologies, we should aim to mitigate the risks of religion while pursuing its opportunities.

Religion already isn't benign, and any religion worthy of a superintelligence certainly would be even less so.
Of course, humans already live in a world of intelligent machines. In our everyday lives, neural networks process information and solve problems and improve while doing it. Recently, they have been improving quite rapidly. And as they're learning to think, their thinking is notably different from humans.

As science writer David Berreby recently reported:
Some hard problems make neural nets respond in ways that aren't understandable. Neural nets execute algorithms -- a set of instructions for completing a task. Algorithms, of course, are written by human beings. Yet neural nets sometimes come out with answers that are downright weird: not right, but also not wrong in a way that people can grasp. Instead, the answers sound like something an extraterrestrial might come up with.

These oddball results are rare. But they aren't just random glitches. Researchers have recently devised reliable ways to make neural nets produce such eerily inhuman judgments. That suggests humanity shouldn't assume our machines think as we do. Neural nets sometimes think differently. And we don't really know how or why.
Thinking machines are different. They will be different. What will that mean when it comes to questions of ultimate things?