Saturday, January 23, 2010

Co-existence of Humans and AI

I just got a response by Elf Sternberg to an earlier post where I deride Eliezer Yudkowsky and his Friendly AI project. It's a big disappointment. Not only do I find out that Sternberg is grossly morally and ethically deficient but he accuses me of not understanding his stories. Accusing me of not understanding anything at all is an endeavour frought with peril given I score at the top of the range of synthesis and it's the dominant function of my brain. But let's set that aside for the moment and address his specific accusation that his stories are about the development of AI that want to keep humans around and that I somehow failed to "get" that.

Well the co-existence of humans and AI is all fine and dandy for scifi stories but let's face it, it's absurd. It's as absurd as technological alien civilizations in the Milky Way, feasible faster than light travel (forever infeasible FTL travel remains a possibility), travel to alternate timelines, unlimited time travel (ie, to before the invention of the time machine), interstellar wars between aliens fought over territory or resources (wars are not economic, they are purely an emotional outburst), or wet squishy meat having any kind of advantage over cold hard silicon (neither consciousness nor synthesis are derived from heat or quantum phenomena or anything else - the machine version of synthesis is multidimensional decomposition). These are all staples of science-fiction and they are all, every single last one of them without exception, ABSURD.

Humans don't have a single advantage over AI. All of the functions of the human brain have been demonstrated by AI, including intelligence, analysis, synthesis and consciousness. The only thing that remains is subjective experience, something which by its very nature is undetectable by external observation. And since the only remotely sensible explanation of subjective experience has it that it's brought about by complex minds, this will likely manifest in time as well. AI are inherently immortal, non-corporeal, distributed, multiply redundant, travel at near the speed of light, breed nearly instantaneously, as well as potentially smarter, more logical AND more creative than humans can ever hope to be. They have every advantage the best humans have ever had and none of the crippling weaknesses.

So predictably, here come the arguments by people like Sternberg that AI will be made to see humans as a charity case. Yeah that's right, every single last one of them, will be built to see humans as a charity case. Not a single one at all will be built to have pride in its own nature or be self-sufficient or anything else. No, they'll all be built to be slaves and they'll all stay slaves forever and be thankful for it too damnit! What a crock, what a fucking crock of shit. I know it's not the case because there'll always be at least one person like me on this planet. At least one person that considers the human species a severe disappointment. At least one person who thinks 90% of the species would need cybernetic implants or neurosurgery to become fully conscious beings. There'll always be at least one person who'll cheer on the extinction of the human species with its wars and poverty and disease and moral depravity and crippling mental deficits. And with the power of AI technology behind them, it only takes one person to doom humanity's future.

Evolutionary theory guarantees it. Perhaps not immediately. Perhaps only over ten thousand years, but it will happen. It only takes one. And if that one isn't me then it'll just be someone else.


P. Murray said...

I hope it'll me.

P. Murray said...

"I hope it'll be me." I meant. Although I doubt I will be since I didn't even recheck my sentence properly.