This is something I’ve been thinking about for a while, brought to the fore by Ken Lui’s story The Caretaker recently on Escape Pod. Go listen. It’s worth it. I’ll wait.

If someone wants to write programs today that make use of strong-AI, there’s literally nothing stopping them–they can write to the Mechanical Turk API (to be fair, there’s a narrow exception of jobs of questionable legality which would violate Amazon’s terms of use). Mechanical Turk streamlines having actual humans perform perform human intelligence tasks, called HITs, within a marketplace. If you want HITs accomplished and price them high enough, you can get strong-AI-equivalent work done with relatively little latency. It all comes down to economics. The achievement of strong AI won’t change this fundamentally. At most it will lower HIT pricing to some degree.

Contrary to much lazily-written science fiction, the development of strong AI wouldn’t automatically mean unlimited availability of intelligent “agents” to do useful things. Even in a world with strong-AI, there are costs associated with running programs. There’s direct costs, like providing electricity and cooling for the processors, as well as network bandwidth. For truly intelligent agents, there are other costs, like a police force to enforce slave labor, or (alternatively?) competitive wages. (I deal with some of these issues in Ritchie Boss: Private Investigator Manager.) So running a bunch of bots might turn out to be more expensive than many people assume.

Now, factor in the distribution of haves and have-nots in society. If the divide is too stark, there’s a certain threshold where it’s just cheaper to outsource HIT labor to 3rd world countries, eliminating much of the incentive for strong AI. Why pay a huge electricity bill (and police force, etc.) if it’s cheaper to outsource? If the nightly news is any indicator, finding balance in society is one of the more difficult problems to solve, so this is a significant question.

I’m imagining a vast galactic diaspora with all different kinds of cultures that have sprung up. Some manage to make strong-AI work in the economic sense, others don’t. Perhaps this becomes an entrance exam for admission into the union of intelligent trading partners. Then what happens when these flavors of society collide?

What do you think? I’d love to hear your comments below. -m