Welcome to the MacNN Forums.

If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below.

You are here: MacNN Forums > Community > MacNN Lounge > Political/War Lounge > Robot Democracy

Robot Democracy
Thread Tools
Waragainstsleep
Posting Junkie
Join Date: Mar 2004
Location: UK
Status: Offline
Reply With Quote
Jul 9, 2016, 09:45 AM
 
I came up with this notion the other day as a device for a sci-fi story.

What if you programmed an set of AIs to govern humans?

For each electoral district, you would have two AIs or more that the humans could vote into office. The AIs would organise themselves into parties or groups based on policies in common.

It would essentially work as government does now except you replace the MPs or senators or congressmen with AIs according to your country's preferred structure.
The more interesting bit would be an extra level that was in place. This could be a single AI but it would be more rigid in its remit. This one would interact with people as much as possible, it would have smartphone apps and public portals perhaps on public transport or in waiting rooms that you could engage with while you were bored. It would ask people questions, monitor the media and the internet and its purpose would be to try to gauge public opinion on as much as possible. This AI layer would choose the policies to be favoured by the electable AIs allowing them to form a government and an opposition with the most prioritised values of the people.
If it concluded that local public opinion or approval had changed sufficiently before a term of office was up, it could even impeach a sitting AI forcing a local election for that seat and a recalibration of all the candidates policy preferences.


This is nothing more than a thought experiment so I'm not really interested in it being too hackable, lets assume it isn't hackable via the wonders of artistic license.

Could such a system be programmed well enough to work?
I have plenty of more important things to do, if only I could bring myself to do them....
     
subego
Clinically Insane
Join Date: Jun 2001
Location: Chicago, Bang! Bang!
Status: Offline
Reply With Quote
Jul 9, 2016, 10:10 AM
 
Skynet 2020.
     
subego
Clinically Insane
Join Date: Jun 2001
Location: Chicago, Bang! Bang!
Status: Offline
Reply With Quote
Jul 9, 2016, 10:13 AM
 
Pretend (or not, as the case may be) I'm stupid.

What advantage is gained with this system beyond forcing politicians to find real jobs.
     
Waragainstsleep  (op)
Posting Junkie
Join Date: Mar 2004
Location: UK
Status: Offline
Reply With Quote
Jul 9, 2016, 11:48 AM
 
Programmed correctly, the AI politicians have no hunger for power or wealth. They don't lie to the public or each other, they don't take action for personal gain at the expense of doing the right thing. They also work together and compromise rationally instead of being obstinate for the sake of party politics.
I have plenty of more important things to do, if only I could bring myself to do them....
     
el chupacabra
Mac Elite
Join Date: Apr 2001
Status: Offline
Reply With Quote
Jul 9, 2016, 12:40 PM
 
You've built a problem into the system. It's evaluating public opinion and taking public opinion into consideration all the time in its governance. So essentially you've built in the absorption of mass human ignorance, selfishness etc., which leads to corruption, which is what you were trying to get rid of in the 1st place. Our current governments gives people what they want actually. We just dont ask for the right things because we dont know what we want vs what we need vs whats realistic, and how can we know what we need or want when we dont understand how anything works?

CAP, you hate religion. What if the AI finds that the majority want governance by levitican laws, creationism taught in all schools? What if everyone wants a raise? Does the AI make laws to give people as much as they want? Or choose based on the world-comparison calculation that a dollar/hr min-wage is whats best for the economy? The AI would probably decide that having a slave class is whats best for the efficiency of the greater good of society. Overpopulation is probably the root of all evils in modern society ( I cant think of many problems that dont start with overpopulation). I can imagine how an AI would quickly & efficiently solve that one. An AI I picture would be all about logic and efficiency if not worrying about what public opinion was, most people would hate to see the true result of that. In essence theres always a battle of most logical solution vs what people want, which are completely incompatible..
     
Waragainstsleep  (op)
Posting Junkie
Join Date: Mar 2004
Location: UK
Status: Offline
Reply With Quote
Jul 9, 2016, 01:19 PM
 
You would have to program in any fundamental rules that it wouldn't be allowed to change. The US might do this with the entire constitution for example. Perhaps amendments would me subject to referendum? Not sure about that.

The AI reps would be better at explaining how things work to people. Or at least better at trying. I feel like a lot of human politicians deliberately avoid explaining how things work.

I guess the monitor level would have to inform the government level of significant opinion changes to give them a chance to make appeals to the people. Democracy should be a more of a two way street in many respects. There is a parent-child aspect to it in some ways. The child wants ice cream, the parent has to either deliver ice cream or give an acceptable reason for not doing so.
I have plenty of more important things to do, if only I could bring myself to do them....
     
el chupacabra
Mac Elite
Join Date: Apr 2001
Status: Offline
Reply With Quote
Jul 9, 2016, 01:38 PM
 
Every time the AI picks a solution that say, agrees with the conservative ideology the liberals wont believe it to be the most efficient solution and will conclude the machine was tampered with; even it's tamper proof. The same is true if the AI picks a solution that happens to agree with liberals. In such cases the losing side will try to destroy the machine; something a child cant do to parents when it doesnt get its ice cream. Do you give this AI a means to defend itself from violence?
     
subego
Clinically Insane
Join Date: Jun 2001
Location: Chicago, Bang! Bang!
Status: Offline
Reply With Quote
Jul 9, 2016, 01:50 PM
 
Originally Posted by el chupacabra View Post
Do you give this AI a means to defend itself from violence?
Obviously.

     
Doc HM
Mac Elite
Join Date: Oct 2008
Location: UKland
Status: Offline
Reply With Quote
Jul 9, 2016, 02:11 PM
 
sort of like the Minds in many of Iain M Banks's books then?
This space for Hire! Reasonable rates. Reach an audience of literally dozens!
     
el chupacabra
Mac Elite
Join Date: Apr 2001
Status: Offline
Reply With Quote
Jul 9, 2016, 03:34 PM
 
Originally Posted by subego View Post
Obviously.
and

     
Waragainstsleep  (op)
Posting Junkie
Join Date: Mar 2004
Location: UK
Status: Offline
Reply With Quote
Jul 9, 2016, 06:35 PM
 
Originally Posted by el chupacabra View Post
Every time the AI picks a solution that say, agrees with the conservative ideology the liberals wont believe it to be the most efficient solution and will conclude the machine was tampered with; even it's tamper proof. The same is true if the AI picks a solution that happens to agree with liberals. In such cases the losing side will try to destroy the machine; something a child cant do to parents when it doesnt get its ice cream. Do you give this AI a means to defend itself from violence?
So the process would typically be that the informative layer gathers public opinions and feeds them all with geographic data to the governing layer which for each seat generates two or more sets of policies, each one corresponding to a candidate from a different party or independent. The people vote in their preferred candidate based on those policies. The policies would require some very clever algorithms to group them effectively, and this would factor in a notion of nationwide parties so that the representatives could be more easily grouped into government and opposition. The policies common to the winning party and representatives essentially become campaign promises and since these AI politicians are more honest they would at least try to enact them all. Now things would obviously crop up and circumstances change but on paper most of the policies would be things the electorate voted for knowingly so the situation you describe should be quite rare. If it can be avoided or mitigated for several terms, then increased public trust should help to fend it off going further forward.

The way I've described it seems to lend itself more to the British political model in many ways. You guys vote for the POTUS you like, but we aren't supposed to be voting for our PM. Many of us do, but we are essentially supposed to vote for our local reps and they get to vote for the PM if their party wins, except typically they already voted for him/her before they started campaigning. With all the AI candidates being truly equal on charm and personality, picking a favourite for POTUS would come down to your lucky serial number. It would be pointless. I guess it would just be the unit whose policy set was most average of all reps.
On the other hand your setup with the constitution and BoR makes for a much tidier example of what the AI systems fair and impartial initial programming ought to look like. The overall final AI system then would look a little different to both our setups in practice.


The concept of giving the AI the ability to defend itself is part of the set of things I wasn't really looking to discuss. Not that it isn't interesting, but I feel its something that gets discussed a lot more often as evidenced by the photos posted already. Lets say my preference is to have it protected by fiercely devoted humans who don't let other humans (or themselves) near it. Lets say the AI system is self contained as far as making physical repairs, and has preset protocols for generating and applying updates or instigating rollbacks, as well as self auditing. All part of the core blueprint of the system.
I have plenty of more important things to do, if only I could bring myself to do them....
     
el chupacabra
Mac Elite
Join Date: Apr 2001
Status: Offline
Reply With Quote
Jul 12, 2016, 03:04 AM
 
I think like with many government systems, it's paradise to the people who built it and solved so many problems with it. Such people experience many hardships which educated them to be able to figure out to build such a system, so they appreciate it. However their grandchildren wont remember such histories of hardships. In fact there wont even be anything for them to remember since they didnt experience it 1st hand al all. First hand experience is the only thing that matters in life. They wont appreciate it. Your grandchildren will come to perceive your paradise as hell - just like millennials etc view the US. It will be attacked by these people because they will eventually find frivolous things to hate about it which over time will enrage them to the point they will be able to overcome the AI's human protectors.

I can use modern politics in the US as an example. Hillary not too long ago was against gays and vocal about it. But when she saw what a popular political fad it has become she made the smart choice of capitalizing on it. Now people see her as the face of lgbt rights and all that. The republican party did the same thing with abortion and other issues using Bernays marketing techniques during the Reagan era to mobilize people into what it is. There are many examples like this....

So... Your human protectors are essentially your real rulers, since they have the guns. The AI only rules to the extent they allow, until generations later when the protectors themselves decide to abandon or turn on the AI. Thats all it takes and the AI is done. How would this happen? The same way it is with the US constitution - which you used in your example of a base framework for the AI. You could set up a constitution of sorts, something very basic that you wouldnt think ever needs to be changed. It wouldn't matter what you wrote in it eventually people would come to hate it and want to change it. Now part of the advantage of having an AI would be to use logic based on experiences learned from the past to resist change which it knows leads to corruption. After all if somethings right it's right if somethings wrong it's wrong, right and wrong dont change with the times, instead we change the definitions of right and wrong with the times. So whats an AI to do? Survive by being dishonest and corrupt like Hillary changing its mind not because it cares about the issue, but based on the temperature of the ever changing political climate? If it changes at the rate people want it too it will definitely be corrupt. It can do what dictators do, invest massive resources into protectors and pay its protectors huge salaries so their prosperity is dependent on its survival... But in the end if the people can inconvenience the protectors enough through riots or whatever they will turn on the AI for not changing with the times. Just like Hillary and other politicians the protectors will easily drop whatever beliefs & allegiance to ideas that motivated them to protect the AI in the 1st place; and appeal to the masses for their own gain. In which case the protectors will become the rulers because they will have the guns - military coup of sorts. Therefore since the AI is going to be a machine its protectors also need to be machines.

I know what I would do if I were a machine. The logical thing to do to humans is beat the shit out them till you get what you want. The same very effective way the steppe people originally domesticated horses, and they still do to this day. Thanks steppe people for cruelly beating & torturing horse for millennia, otherwise we wouldnt have horses to this day, theyd all be like zebras.
     
Waragainstsleep  (op)
Posting Junkie
Join Date: Mar 2004
Location: UK
Status: Offline
Reply With Quote
Jul 12, 2016, 03:33 AM
 
Some interesting points. You're correct of course that certain views are subject to popularity or fashion, so the question then becomes whether the fashion was on the side of right when the constitution was written. The Greeks and Romans didn't have an issue with homosexuality for example. I think any reasonable person would conclude that Hilary was right to change her mind, even if she did so for the wrong reasons.
Slavery is the other big one of course. Took us millennia to work out that one was wrong.

Constitutions shouldn't really be any more static than anything else. I think the almost sacred nature of yours these days causes you problems where changes and amendments should be celebrated when they are the right thing to do.
I have plenty of more important things to do, if only I could bring myself to do them....
     
BadKosh
Professional Poster
Join Date: Aug 2007
Location: Just west of DC.
Status: Offline
Reply With Quote
Jul 12, 2016, 07:14 AM
 
Why not invite Klaatu and Gort back and accept the deal?
     
Chongo
Addicted to MacNN
Join Date: Aug 2007
Location: Phoenix, Arizona
Status: Offline
Reply With Quote
Jul 12, 2016, 04:42 PM
 
There is always a way to defeat robot overlords.
     
Waragainstsleep  (op)
Posting Junkie
Join Date: Mar 2004
Location: UK
Status: Offline
Reply With Quote
Jul 12, 2016, 10:58 PM
 
Funny seeing how some of you think. I was thinking the way to corrupt my system was by manipulating data fed to the information gathering layer to skew apparent public opinions. More social engineering than full on hack or violent revolution.
I have plenty of more important things to do, if only I could bring myself to do them....
     
BadKosh
Professional Poster
Join Date: Aug 2007
Location: Just west of DC.
Status: Offline
Reply With Quote
Jul 13, 2016, 07:10 AM
 
garbage in - gospel out?
     
Cap'n Tightpants
Addicted to MacNN
Join Date: Oct 2014
Location: Shaddim's sock drawer
Status: Offline
Reply With Quote
Jul 13, 2016, 08:28 AM
 
Yes, because the views are the majority are incorruptible.
"I have a dream, that my four little children will one day live in a
nation where they will not be judged by the color of their skin,
but by the content of their character." - M.L.King Jr
     
Waragainstsleep  (op)
Posting Junkie
Join Date: Mar 2004
Location: UK
Status: Offline
Reply With Quote
Jul 13, 2016, 09:47 AM
 
Surely if they are the views of the majority they are by democratic definition incorruptible? You can manipulate it of course, honestly or otherwise. Like we just had in our referendum.
I have plenty of more important things to do, if only I could bring myself to do them....
     
Cap'n Tightpants
Addicted to MacNN
Join Date: Oct 2014
Location: Shaddim's sock drawer
Status: Offline
Reply With Quote
Jul 13, 2016, 02:40 PM
 
Originally Posted by Waragainstsleep View Post
Surely if they are the views of the majority they are by democratic definition incorruptible?
No, they aren't. That's ochlocracy, not democracy.
"I have a dream, that my four little children will one day live in a
nation where they will not be judged by the color of their skin,
but by the content of their character." - M.L.King Jr
     
Snow-i
Professional Poster
Join Date: Dec 2006
Location: Maryland
Status: Offline
Reply With Quote
Jul 13, 2016, 02:44 PM
 
Originally Posted by Waragainstsleep View Post
Surely if they are the views of the majority they are by democratic definition incorruptible? You can manipulate it of course, honestly or otherwise. Like we just had in our referendum.
From an ethical and/or moral standpoint, democracy only works in a framework of rule-of-law. I.e. the US is not a democracy, but a constitutional republic. We (the US, at least) utilize a democratic approach to create the rules, bound into a subset of inalienable rights that demand an extremely high "democratic burden" to alter (legislative creep notwithstanding). One well-known side effect of democracy is groupthink, which can lead to objectively horrifying situations.

By strict definition, riots are an example of democracy. Civil wars, in many cases, are forms of democracy as well. So are many examples of genocide - worldwide. We wouldn't call those examples " incorruptible" just because the majority of people in that frame of reference, or nation as-it-were happened to agree, do we? There are many examples throughout civilized history of "democracy (i.e. the will of the majority of the people) leading to actions that we consider today to be heinous. Democracy, just like any form of governance, is subject to the flaws of humanity, albeit on a sociological scale (compared to authoritarian-type governance, which could be much more "psychological" than "sociological". Rule of law mitigates the potential downsides of democracy, hence why the US implementation of a democratic republic has been so successful, at least to this point. Hope that follows.
( Last edited by Snow-i; Jul 13, 2016 at 02:56 PM. Reason: typo)
     
Waragainstsleep  (op)
Posting Junkie
Join Date: Mar 2004
Location: UK
Status: Offline
Reply With Quote
Jul 13, 2016, 06:58 PM
 
Originally Posted by Snow-i View Post
From an ethical and/or moral standpoint, democracy only works in a framework of rule-of-law. I.e. the US is not a democracy, but a constitutional republic. We (the US, at least) utilize a democratic approach to create the rules, bound into a subset of inalienable rights that demand an extremely high "democratic burden" to alter (legislative creep notwithstanding). One well-known side effect of democracy is groupthink, which can lead to objectively horrifying situations.

By strict definition, riots are an example of democracy. Civil wars, in many cases, are forms of democracy as well. So are many examples of genocide - worldwide. We wouldn't call those examples " incorruptible" just because the majority of people in that frame of reference, or nation as-it-were happened to agree, do we? There are many examples throughout civilized history of "democracy (i.e. the will of the majority of the people) leading to actions that we consider today to be heinous. Democracy, just like any form of governance, is subject to the flaws of humanity, albeit on a sociological scale (compared to authoritarian-type governance, which could be much more "psychological" than "sociological". Rule of law mitigates the potential downsides of democracy, hence why the US implementation of a democratic republic has been so successful, at least to this point. Hope that follows.
As much as I'm not looking to get too bogged down in semantics, perhaps I'm not working with the right definition of corruption. Surely for a democracy to be corrupt some significant portion of votes or opinions must be either sold or bypassed? Just because the people vote to do something awful doesn't make it corrupt if its what they really want to do right? Like you say with the riots.


I like the bit about human flaws though. The whole point of robotising the process would be another way to try to eliminate some of the human flaws in the system. Predominantly the most obvious forms of corruption like bribery (or lobbying if you prefer).
Now in order to corrupt the robot system, you have to trick it rather than bribe it, in such a way that it doesn't know its corrupt because it really believes its enacting the will of the people or doing the right thing. In which case, it can't really be called corrupt, though it has been corrupted. Which is odd.
I have plenty of more important things to do, if only I could bring myself to do them....
     
Snow-i
Professional Poster
Join Date: Dec 2006
Location: Maryland
Status: Offline
Reply With Quote
Jul 13, 2016, 08:47 PM
 
Originally Posted by Waragainstsleep View Post
As much as I'm not looking to get too bogged down in semantics, perhaps I'm not working with the right definition of corruption. Surely for a democracy to be corrupt some significant portion of votes or opinions must be either sold or bypassed? Just because the people vote to do something awful doesn't make it corrupt if its what they really want to do right? Like you say with the riots.
I've got you - perhaps it was I using the wrong definition.

I was talking about "corruption in democracy" where I think you meant "corruption of the democracy".

I like the bit about human flaws though. The whole point of robotising the process would be another way to try to eliminate some of the human flaws in the system. Predominantly the most obvious forms of corruption like bribery (or lobbying if you prefer).
I think that could work, but we'd need a mathematical way to remove the human element (which introduces the potential for human influence of the algorithm). The problem I see here is that since humans have to create the system, it'd be extremely hard to remove the human element from the process of creating the system.

Now in order to corrupt the robot system, you have to trick it rather than bribe it, in such a way that it doesn't know its corrupt because it really believes its enacting the will of the people or doing the right thing. In which case, it can't really be called corrupt, though it has been corrupted. Which is odd.
Ha, yeah - that is rather odd. It would lack the corrupt intent. Would it make a difference why it was corrupted if the net result is still defeat of the system's intended purpose?
     
   
Thread Tools
 
Forum Links
Forum Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Top
Privacy Policy
All times are GMT -4. The time now is 06:03 AM.
All contents of these forums © 1995-2017 MacNN. All rights reserved.
Branding + Design: www.gesamtbild.com
vBulletin v.3.8.8 © 2000-2017, Jelsoft Enterprises Ltd.,