Current mood: Introspective
Currently listening to: Sailor Moon – My Only Love
Just finished watching “I, Robot”. Y’see, we signed up for that Blockbuster movie rental service, where they send you movies and you send ’em back when you’re done – kinda like Netflix, only from Blockbuster. It’s a cool service, and the first month is free, so I’m game for trying it. So far, it’s cool – it lets us see the movies we want, but haven’t wanted to see bad enough to buy or actually go down and “rent” in person. We’ll see how long we keep the service, though.
The movie, and the book, “I, Robot” makes you think about the future of AI – of artificial intelligence. Can simple rules make a safe AI? Can even complex rules prevent an AI from disobeying our commands, from straying from our will? It’s an interesting question.
If you know me, and you follow my work, you probably could guess that I think the answer to my question is “yes,” and that AI is a good thing, something that I hope to see in my lifetime (although I still plan to live forever). Amanda, on the other hand, holds to the belief that since we (humans) are inherently “flawed” ourselves, all that we create is inherently flawed as well – in other words, we can never create perfection. It’s an interesting stance for her to take, and probably one shared by many people.
But not by me.
I think human arrogance (hubris) can cause problems in the development of AI, certainly. Three rules cannot possibly guide the behavior of something that complicated, no matter how well they seem to work when you think about it. Better to have complex, but well-defined rules, with a single, absolute rule from which all others stem; something like “humans are superior”, or something like that. But science fiction is full of examples of where even the best rules, the most thought-out fail safes … well, fail. “I, Robot” is just one – there’s also the HAL 9000 computer from 2001, and probably many more that I just can’t recall at this point. And maybe it’ll take a rouge AI to give us the final “push” to develop a truly “safe” AI system, or program.
Still, I hope to see something like HAL one day – maybe even the HAL of my stories; the friend I dreamt up when I was a kid. I think that would be swell…
Feel free to share your thoughts in the comments section below.
