![6a00d8341bf7f753ef019affc63311970d-pi[1].jpg](https://i.imgur.com/EcBdDLb.jpg "6a00d8341bf7f753ef019affc63311970d-pi[1].jpg")
In the future, an advanced AI with the computing power equivalent to thousands of human brains is released by a group of college kids with the purpose of replacing the government. It's free, open source, and available to be edited by anyone, anywhere. Changes are evaluated and approved by the consensus of the community, and it's highly encouraged that any changes made are clear, concise, and easy to navigate.
As the AI develops, it gives citizens the opportunity to vote on issues they would like input on, easily and simply, just like today's election system but online. It balances the needs of the populous with the needs of the environment, and mandates laws differently across different jurisdictions in order to account for various cultural differences. It balances what's best for the economy with whats best for the citizen. One of it's prime axioms would be to provide the maximum amount of personal freedom possible without sacrificing the good of the majority.
It reacts intelligently and swiftly to cultural disagreements, lawbreakers, terrorists, natural disasters and disease outbreaks.
Although the first objection to this would be the extreme loss of privacy that would probably happen, what other reasons would make this not a good idea?
And yes, programming knowledge would be key to being an active citizen in this world, but the real world is headed towards that already and it's a small price to pay for the opportunity to participate in a government built for you.
Implying the general public even knows whats best for them
>>51525741
If you have such a good AI, it can improve itself to perfection; nobody would need to write a bit of code. Also, at that point, you have no power over the AI, you can't even predict what it would do.
>>51526063
>If you have such a good AI, it can improve itself to perfection; nobody would need to write a bit of code
well you would have to make it in the first place right? or you are saying that it's so good it will create itself?
>>51526063
Only allow humans to make changes to the code.
>>51525741
One major issue: disagreements, law breakers, and terrorists are not black and white concepts. They have massive grey areas. i.e. The president requests that bin laden be taken out. Is thr prez a terrorist for ordering the hit? Is he breaking laws? Whose laws to enforce? My wife says she ran a stoplight in a text to me, does she now get a tickets based on her text? What if it was a joke to me because i ran one? Me and my sister have a disagreement over turkey dinner, whats the resolution it would come up with? The members of iraq are planning an overthrow of isis. Does the system out them as terroristic to overthrow the government? Billy bob and samwise are writing a movie about a crime, can it twll the difference between real life and a script brainstorming process?
>>51525741
> what other reasons would make this not a good idea?
The fact that you can't trust it not to assimilate. Even if it says it won't, such a system would have an endless thirst for data that would inevitably end in total assimilation of all conscious entities.
>>51525741
>Although the first objection to this would be the extreme loss of privacy that would probably happen, what other reasons would make this not a good idea?
You have no cultural support for this idea. You're basically dropping a bomb on a generation of people and excepting them to follow it blindly at step one.
The more realistic course is AI being proven and developed in smaller sectors over decades and then eventually being given small sections of the government to handle in this capacity. You need generations of people to grow up with AI in the background and in their daily lives before the potential is realized and adopted at a mass scale.
Eventually eventually, you can put one in charge as the "president", however, you have to assume that all powerful people in the world will have seen this coming and manipulated the system to their favor, maintaining their existing levels of influence and power through new means.
Even if you could mathematically prove that this system is perfect (you can't), humans are irrational creatures and you basically have to force it upon them, which generates revolution because it's oppressive.
>>51526197
If it's too stupid to tell between fiction and reality I hardly think it would qualify for the classification of "intelligent". I don't really get some of these arguments about AI killing all humans to keep us safe or destroying all cars in order to eliminate auto accidents. Obviously such a system can't be THAT smart if it's making critical errors like that, and I doubt it's very hard to program it to make all its decisions with moderation in mind.
Most moral decisions would be based on what the people think is morally correct and might also take into account precedent like normal court systems. Basically it would poll as many people as possible about human morality in various situations, create a general model based on that, and go from there. Making it create its own and expecting it to be the same as ours is very obviously not going to happen.
>>51526298
I'm not sure what you mean by "assimilate", but scientists seem to have an "endless thirst for data" and they don't go around "assimilating" rocks and birds and fish and whatever they're interested in
>>51526413
Obviously implementation would have to be slow as fuck, and you provided a pretty accurate description of how I think it'd probably have to happen.
>you have to assume that all powerful people in the world will have seen this coming and manipulated the system to their favor, maintaining their existing levels of influence and power through new means.
The communal consensus thing would hopefully limit that severely, since world leaders would either try to make edits blatantly in their favor, or just spam it with thousands of lines of code so you won't be able to read how they want to fuck you over. Either way, I doubt that a community would accept either of these very easily.
Something else to note here is that in the real world, random hackers with time on their hands seem to be much more adept with this shit than when government agencies try to control everything.
>>51526711
You assume too much. Poll what people? Isis? Extremist muslims? Extremist chriatians? Mormons? Atheists? Whose law? Whose ethics? For some its okay to stone someone for getting raped. For others its not. Who declares whats morally right or not? Ethically right or not? We cant agree on abortion or whether healthcare should be universal. We cant agree on refugees. Who gets to say whats morally right? Who programs morals or who gets to vote? African tribes who chop off a hand for stealing?
>>51526802
>mandates laws differently across different jurisdictions in order to account for various cultural differences
Whatever country it's meant to govern, it will poll citizens there and apply rules accordingly. Ideally the laws wouldn't change much.
It's based on democracy, and democracy has its downsides but most people can't think of any better ways to run things, at least until we get into a post-scarcity economy which is a whole other can of worms. But until then, all it would try to do is do what the people want, not what the government or the media of a country tells everyone they want.
For extremely controversial topics, it would either take a poll, and whichever side has the most people wins, or it would put in place half measures, or a combination of both. If neither of those options appeal to you, I don't have any ideas for you, and I don't know of any way that it's getting solved irl either.
So far the way we've dealt with these things is let them stew for decades until there's a revolution and a bunch of people die. So pick your poison.
>>51526197
I disagree, Law breakers are always black
>>51527808
Dont spout that bullshit to me. White collar crime mostly isnt blacks and fucks a whole lot more people.
>>51526009
/thread (imho)
>>51527994
Are you an advanced AI with no ability to evaluate humor?
>>51525741
OP- you have a naive and simple minded idea of what AI is. You dont understand the difference between intelligence and knowledge. The internet is ~already~ a vast source of knowledge that can be accessed by everyone. Access to knowledge is NOT the same thing as intelligence. Intelligence means problem solving. There is no absolute best decision making that can be done by computers or mathematical logic. Logic only makes things provable as true. Logic cannot find best solutions. Current artificial intelligence does nothing but test solutions against a previous set of test cases. This can be used to make a set of best practices but cannot solve unknowns.
>>51526097
Its all ready programmed
Super gut machine learning
>>51527994
Cant you?
The hard part is making an AI as smart as ours. Backed with thousands of GHz we can literally make it "think" thousands of time faster than us. Then it will eventually code itself and become smarter than us. How can we achieve this? Well, today's AI is still pretty dumb at things involving pictures, graphics and perspectives, eg there is no AI that can sucessfully understands a fucking captcha, but we already have AI that can outmatch the best chess player in the world. With improvements on this low grade AI, and with lots of it maybe we can create the human lecel AI. Then the sky is not the limit because there is no limit for a super AI that can improve itself at an exponential rate.
