Compromise
Maybe an A.I. could monitor/guide certain sectors of the economy to benefit humanity without bias or ulterior motives. Would certainly end this capitalism vs. something-other-than-capitalism stuff. Any thoughts?
posted by Mookie at 5:33 PM
7 Comments:
I had never thought of this before, so I slept on it last night.
What makes you think that an AI could be without bias? I do not think it could.
After all, what is intelligence besides being able to respond to stimuli in a certain way? Of course, for true "intelligence" and not simply programming, previous stimuli (events) must have an affect on what the outcome is for later stimuli. A computer programmer would not be able to say "If A happens, run B" because the AI must be able to adapt in case it is shown that in previous instances action B is undesirable. My whole point being that for true AI, the AI would be a product of its previous experiences, i.e. environment. I think it would be biased towards that environment.
delta,
"I think it would be biased towards that environment."
Good point you make, and well worth the concern. Could be a matter of some tweaking to smooth it out.
I would imagine the AI would have some specific guidelines and goals to help it direct affairs. The affairs may change, but the ideals do not. And yes, the AI may get stuck in one "mindset" regarding a current state of affairs, but here is where the helper monkeys come in to readjust or reset the "mindset". Besides, I don't know how far off base the AI would get from reality, given an almost constant flow of data. I think the AI would help stabilize the system, and even if the system were to change, it would not do so drastically.
I would imagine the AI would have some specific guidelines and goals to help it direct affairs. The affairs may change, but the ideals do not
If it had guidelines and goals, then could it truly be AI, or would it just be some highly elaborate program? And who would decide what those ideals would be? I'm assuming they would be decided by democratic means. But if many people agree on the ideals that the economy should represent, why couldn't we just have humans carrying out the necessary calculations on ordinary processors?
And if the ideals were democratically decided, wouldn't it be preferable to have a non-capitalist society first, before the ideals are decided? This is so that people have time to build solidarity with one another. If ideals were voted on today, one of the top 3 would be "devotion to God".
"could it truly be AI, or would it just be some highly elaborate program?"
It would received updates on progress and such, so it could adjust accordingly.
See: http://archives.cnn.com/2000/TECH/space/12/27/part.four/index.html
"ouldn't it be preferable to have a non-capitalist society first, before the ideals are decided?"
I don't know if humans are trustworthy enough. An AI would not care if it didn't get a $398 million severage package upon retiring. What will likely happen is it will be a bit of both to acheive the aim of both. The AI can help stabilize the economy, help with unemployments, etc, or whatever else its designed to do, while the humans try their best not to be idiots.
That's an interesting article, although I don't think I would consider it truly AI. If it simply looks at the environment and then decides which action to take (based on its programming) then it is not AI imo. For true AI it had to be able to develop new ways of thinking about a problem, to create new modes of thought. This would require that it be able to program itself to execute the new ideas.
And I still don't see why AI would be useful. If humans are going to tweak it whenever it doesn't do what they want, what good is it? "AI" is useful in terms of space exploration because of communication problems with the spacecraft, but there would be no such problems on earth.
I also don't see how it would make irrelevant the capitalist vs non-capitalist debate. If the AI runs the entire economy, then it is a centrally planned economy. If it is run to maximize the needs of the people, then it is socialist. If it is run to maximize profit for the individual owners, it is capitalist. Either society will keep privelege and power or it will get rid of it.
Not to mention that if AI was developed in a capitalist society, mechanized police and military would make any sort of social change impossible to achieve if it went against the interests of the ruling class.
Well, a friend of mine is a mathemetician, and he knows of a job position in some companies that calls for C/B analysis. They work on equations that maximize profit for the company. I know of a good example of this. A few years ago an album I was hoping would come out in the US was delayed because it was to compete with some pop singer. The record company held off the release date to make sure the wave of popularity for the pop artist waned. (I just ended up downloading the album and have not paid for a CD since.) A building in downtown Austin has a computer automatically managing many of its functions; AC, security, water, etc. Another example is WalMart. They have a sophisticated inventory system that tracks sales of particular items, which, as we know, are sold all over the country, sometimes all over the world. The demand for the product can be seen in how many units sell in how much time. Wal Mart takes these numbers and adjusts the price accordingly, to maximize profit. This led me to believe that AI would soon make such decisions.
I then imagined if it would be good for the economy. It wouldn't need to be centrally planned, nor entirely planned. Just a few key economic valves/levers could be run by the AI. Interest rates, bonds, prices, etc. could be handled in part by the AI. Also, it could be based on distributed computing, a lot like SETI@Home and Folding@Home work now. Maybe each city has a regional AI that helps it manage its local economy, and they link up to facilitate trade between regions. There are lots of possibilities for computers to aid in the economy.
"If it is run to maximize the needs of the people, then it is socialist. If it is run to maximize profit for the individual owners, it is capitalist. Either society will keep privelege and power or it will get rid of it."
I did another post kind of like this one where I elaborate a bit more on this idea. Maybe it makes more sense because it has more detail.
"Not to mention that if AI was developed in a capitalist society, mechanized police and military would make any sort of social change impossible to achieve if it went against the interests of the ruling class."
I realize this is all speculation mingled with a touch of idealism, but I see it as something that is likely to happen, perhaps in the near future. Not for the benefit of all, of course, but for the corporations. Wal-Mart will be one of the first to have one installed. I agree that those who have the reins now will be unwilling to give them up, least of all to a computer. But if people did decide they wanted to move towards a more equitable economic system, but couldn't trust businesses or the state to accomplish this, one of the few resorts left is very advanced computers aiding humans in their endeavours.
I don't know if you play computer games, there's a really good civ-style game that has a lot of these ideas: Alpha Centauri. I highly recommend it, even if you are not an avid gamer. The memes in the game are well worth the look. Anyway, in the game, this particular tech is called "Sentient Econometrics" and is described as "using machines to predict the market". This idea is a bit far from what we are capable of right now, but something similar is certainly in the works.
"I also don't see how it would make irrelevant the capitalist vs non-capitalist debate."
The claim of many a capitalist is that an economy cannot be managed because it is too chaotic. I never believed that, and I think a powerful enough computer/AI could do it. They could not argue with an AI that didn't care about material gain, as they would for any public entity that sought to curb their activities, i.e., the state.
Thanks for the replies. Ideas need to be tossed around.
Perhaps I'm being too strict with my definition of AI. It seems that you're talking about it in terms of current technology similar to those buildings or the space probe, whereas I was thinking more along the lines of artificial intelligence, as in the kind that you see in movies about the future. Robots that could truly think for themselves, perhaps develop emotions, etc.
But I agree that in the future, whether it be a socialist or capitalist economy, computers will play a large role. The type of AI that you're talking about could take a great deal of work and human error out of any process that it was involved in.
It's intersting that all of the examples you mentioned I am familiar with. I used to run Seti@home and einstein@home on my computer and ran folding@home on my fiancee's. I also played Alpha Centauri, although it's been a long time. That's a very old game. I'd always attack the religious group first =)
Post a Comment
<< Home