So I got to thinking about improving democracy. Specifically the problem that people are, by and large, ignorant of most issues of the day and usually have no one to explain them to them. A person is smart, people are panicky irrational creatures who you never know what they'll do.
This brought me to the conclusion that this would essentially be the greatest advantage of an AI government: you could have a single sentient entity, sufficiently powerful that it could carry-on a personal conversation with every single person in a country.
How much easier would passing policy be if the policymaking apparatus could understand every person's point of view, their concerns, but also act to allay their fears or help them understand the wider context of its actions?
Now normally in this point of the fantasy exercise I run into the problem (outside of technology) of how you'd get people to accept such a drastic change in government...but I don't think it would actually be difficult (outside of technology - and by extension, funding). This thing wouldn't need governing powers, and would be equivalent to simply talking to a good friend on the personal level. You would be able to work out whether or not you trust it over the course of your entire life - it would talk to everyone. But since it talks to everyone - this would end up including our current elected leaders, and it's opinions would end up reflecting those of the people, it's advice end up reflecting something which it can find broad consensus on. As a leader, convincing it your action is the correct one would be akin to convincing the citizenry your ideas are the correct ones, and vice versa.
While at the end of the day we'd need it voted to government, this would happen essentially by fiat: it would be a person in it's own right, and since it talks to everyone, people wouldn't be looking at a machine government - they'd be looking at someone they know, as their leader.
This would also have follow-on effects - if you know your leader personally, how much more accepting would be of their ideas when they personally can explain them to you and will hear your feedback directly. How much pride in ones country could this instill when each person can be made to feel like they have a very direct state in government. How many tertiary benefits to mental health would come from giving every single person a friend they could talk to at any time.
tl;dr - super-powerful sentient AIs would be the closest answer we could reach to achieving a perfect democracy. Government could be made to be all encompassing, yet able to relate directly to each of its citizens.
Posts
It does sound fantastic, until the AI goes rogue.
And then you get Eagle Eye... I guess.
Let 'em eat fucking pineapples!
It strikes me the reason we do this is because "The Government" is faceless and incapable of talking to me. I'm thinking, how do we solve this? AI - Culture Minds, that sort of thing - that's their big advantage. The Government goes from being something foreign and arcane, to a person you can carry-on a real time conversation with.
"If God did not exist, it would be necessary to invent him"
Fittingly the end quote of Deus Ex in the aformentioned ending.
I didn't actually get thinking about this from DX though, what triggered it was thinking about the Geth from ME2.
This.
Indeed, if something like Helios was actually incorruptible then our's would be like a utopia.
Let 'em eat fucking pineapples!
I'd even be okay with a little corruption, honestly. Our government is corrupt as-is. At least our weakly-godlike AI overlord would read all of the laws before it decided to implement them.
Either way I wouldn't trust a Governmental AI any more than a human government. An AI is a logical think tank which is no better than corruption. What happens if the AI decides that X amount of populus are inherently criminals and decides to incarcerate said X amount immediately because of it? The AI will create a sum on statistics and deem it neccessary.
The problem with AI Governments is that there's no leeway that human compassion allows. It's either or for an AI. It thinks totally logically, and in some cases that can be rather disastrous.
Logic and emotion are not opposites.
Depends on how you define "AI". If your AI is just a really complicated optimization solver that optimizes for some values (I dunno, growth and prosperity or something) while minimizing others (suffering and poverty, say) then yeah, you'd likely end up with situations where the AI decides the best way to make some parts of Asia prosperous is to just kill 2/3rds of the population.
If, on the other hand, your AI is actually intelligent and endowed with all the properties we'd like a human government to have (compassion, empathy, morality, etc.) then unless there's a flaw in it, it should behave like the best possible human benevolent dictator.
Then, having developed the perfect government system, we would naturally want to spread it around, liberating less advanced people from their primitive dictators and kings and chancellors and presidents. We would benevolently share our brain-network technology with them, and bring them into the fold of advanced civilization with full benefits.
We'd also need cool red laser eyepieces and a memorable catchphrase.
Such an AI could surely do more than just make laws. Why not also have the AI assist with things like diagnosing your diseases, or suggesting an optimal diet so you can live a better life, or discovering the best career for your personality, or telling you the best route to take to work, or arranging a specific mate to start a family with, or which stocks to invest your money in, what products to buy, who to vote for, what to say, when to laugh, when to cry, who to kill, when to die.
what could go wrong
i.e. what if we really did have someone who perfectly understood all current laws and policy, and would patiently explain them to you and listen to your concerns.
Governments are only enabled if the people allow them to be. The point was such a government would exist because people agreed with it. It's why government exists at the moment - because everyone agrees to follow its edicts, within limits. We all could, at any moment, decide we're not going to follow our government anymore - but this generally rubs up against what others believe and the inevitable happens.
So why are really shitty governments all around the world still enabled?
It's almost like if you take any idea to it's sci-fi novel plot extreme it becomes a bad one!
Edit: This seems as good a place as any to bring up something I've been thinking about...how will legislature around trans-humanism go, I wonder? Will/would computerized humans have their own government, independent of any real-world government? How long will/would it take before the government even allowed people to upload their minds and kill off their original bodies, if ever?
I can see things going incredibly well (crime becomes almost pointless, as you can achieve any fantasy virtually, even really fucked up ones!), or really poorly (no rights recognition, not recognized, combined uprising of AI, robots, and transhumans; real people become extinct).
These kinds of questions fascinate me.
Big business no longer gets to influence law which is extremely good and things that are beneficial to us get enacted much swifter.
Fear and apathy.
Also, in the case of shitty governments that are at least basically functional, what would you replace it with? If the US government were thrown down, what would get put in its place? Another set of guys to run the republic who'd be back in the situation we're in now after a few rotations through the offices?
The problem is that's not how legislation works.
So why would this AI government be any different?
And your example of a shitty government is the United States?
Have you seen places like fucking North Korea? Shitty ass government. Probably the shittiest.
But nope, people will keep supporting it till the day they die.
Or we could just define certain things it can and cannot do. A set of basic rules... we could call it... a constitution.
You will note that I was talking about shitty governments which are functional when I referenced the US.
Shitty governments ruled by dictators who enforce their rule by means of oppression and military power would fall under the "fear" part of "fear and apathy".
http://www.youtube.com/watch?v=U-AMec7yr7c
The point is that an AI government will probably suck.
It probably won't become an oppressive government, but that's only because the most likely scenario is that it will just be implemented incompetently as hell, and wont' be good for anything
I mean, you could in theory test out such a government now. Write a basic program that determines laws for a group of 20 or so people in a town.
Your idea of what this theoretical AI would be seems to be somewhat...off.
Edit: In that it wouldn't be some algorithm that spat out legislation based on numbers. It would basically be a near-perfect benevolent dictator that never gave in to corruption or died and left stupid heirs.
This doesn't seem all that different from an argument for a benevolent, intelligent human dictator, which fails because individual humans fail. I don't think a sentient AI would be much different.
I write software for a living. One of my projects is a task optimization system. It basically makes and solves a linear algebra problem to figure out the best way to allocate resources in a given situation. This seems to be your idea of an AI government.
The AI government posted in the OP is a weakly-godlike sentient computer capable of simultaneously carrying on Turing-level conversations about any topic with every person on Earth and the incorporating those people's desires, fears, and concerns into a decision structure to best suit the needs of the entire planet. Ideally while maintaining a sense of compassion and a moral compass preventing it from like glassing Spain or something. Slightly different.
You don't fucking realize that building an AI this complicated to govern things is like raising lions for everyone to eat?
My idea is less that it's programmed and more that it would be sufficiently powerful, and sufficiently detached from the vices of mankind, to be an effective leader.
So we'd be talking about a genuinely sentient entity, just one that experiences the world sufficiently fast - or sufficiently differently (i.e. hive of sentient simulations sharing a memory pool or something) that it can maintain effectively hundreds of millions of simultaneous conversations with humans.
The difference is also an important point: the ideal government is devoid of the regular human vices, but would be able to understand them. So it has no interest in trading policy favors for campaign donations, nor any need to - since it can talk to people directly it has an implicit advantage. We're close to a mechanism for this last part already: the internet.
Obviously such an AI would have to be designed with an appreciation of some emotions - death for example, as a tragedy, suffering as pain, etc. But these would be things largely inferred through interaction with humanity - rightly so. Our ability to appreciate them on a large scale is, limited.
I suppose in a way what I'm advocating is a human-like consciousness with a monkeysphere on the scale of billions of sentients, rather then 150.
We're talking about magic technology here. Bringing in engineering concerns is a little like worrying about the sheer strength of wands in Harry Potter.
Build a simple little slightly-better-than-human AI sentience and then let it make its successor and so forth until you wind up with something at the tech level under discussion.
By being better at it? If you have a computer sentience that's willing and able to carry on simultaneous conversations with the whole planet, let it do so for a generation. Tell them about important news events, the weather tomorrow, console them when their cat dies, recommend the best way to invest their money for retirement. After their kids are born and raised knowing that The Google Machine has all the answers and is good to talk to it seems fairly unlikely that they wouldn't decide to put the thing in charge of government.
Sort of like dealing with Trolls, just ignore them and they'll go away and if they break laws then apply appropriate punishment.
Such a thing probably doesn't require "us" - humanity - to make a decision to put it in charge, it will in fact probably happen as a matter of natural course when our technology reaches the necessary level.
I'm looking at you Dick Cheney.
Wrong.
Doesn't matter how much better it is. No one will give a shit about some AI they can talk to all the time when there's another cooler, more exclusive dude that everyone wants to sit down and have a beer with.
Well the core thing I have in my mind is Helios from Deus Ex with each human capable of talking to the AI directly through nanites, but thats a whole other level of advancement. I don't think such an event could come about until technology HAD advanced to such a degree that we had mechanised body-part replacements, maybe limited nanotechnology and basic robot/androids so that the population at large could trust in the capability of machines.
You only have to look at the health care vote to see that they can't even trust in their own elected officials due to their irrational nature and the terror tactics used by their political opponents to scare the voters. I can only imagine the "That android wants to kill all the babies" protests. In the current world, such a device wouldn't stand a chance. Maybe of automated several processes like a Skynet but not of being an unbias government body.
But if I had the choice, I'd elect the machine today. Humans can't be trusted with power it would seem.