8
$\begingroup$

In a setting where there is a multispecies Galactic Democracy with Proportional Representation that has expanded with the rule of giving one vote to any individual, how should they deal with a Machine Intelligence that controls an entire solar system and wishes to join? They have already enjoyed two decades of peace and close foreign relations, and both sides want to try to have a fair deal that do not undermine the Galactic Democracy's values.

Is it most likely that the Galactic Democracy would push for:

A. Give the entire AI a single vote.

B. Give the AI a set number of votes greater than 1.

C. Give the AI a number of votes based on a certain criteria (ex. Energy Production, Land Area, etc.)

D. Decide that their society is incompatible with the AI joining and try to pursue closer foreign relations.

$\endgroup$
16
  • 1
    $\begingroup$ Welcome to worldbuilding. As state in our help center, which I advice you to carefully read, we prefer questions which can be answered in a measurable way. Your post it's a whole matter of opinion, depending on the criteria that the power in place deem more suitable. $\endgroup$
    – L.Dutch
    Commented Jul 11, 2021 at 7:28
  • 2
    $\begingroup$ Political settlements are made for many combinations of reasons - e.g. economic, military, existing law, social view points. You have provided none of the information required to determine what is likely. You might consider how e.g. the EU considers prospective new member states as a rough guide to the issues. Also note that partial economic access is possible to such systems. $\endgroup$ Commented Jul 11, 2021 at 7:37
  • 2
    $\begingroup$ What is the current structure of the democracy? Is it a direct democracy of some futuristic kind, or a more modern representative democracy with, e.g., planets electing members to some kind of Galactic Senate? $\endgroup$
    – Cadence
    Commented Jul 11, 2021 at 9:25
  • 3
    $\begingroup$ Does "Machine Intelligence that controls an entire solar system" mean that it's an overlord controlling billions of intelligent slaves? Or does it mean that the entire solar system contains only one intelligent being? $\endgroup$
    – user535733
    Commented Jul 11, 2021 at 17:14
  • 1
    $\begingroup$ There are 11 answers to this question and only 3 upvotes - one of them mine and I did not answer. Peoples of the stack - peoples! if it is good enough to spend time making an answer it is good enough to upvote. I think maybe people got so into the answer they forgot to upvote. Come back and vote. Al needs rep. $\endgroup$
    – Willk
    Commented Jul 11, 2021 at 20:18

12 Answers 12

10
$\begingroup$

It depends on the population, or else the AI would argue based on the size of its "nation".

If the AI governs a population, then the AI would get as many votes as the population it represents. Unless ofcourse the AI tries to cheat by creating many small AI's and presenting them as persons.

If the AI is just itself across a solar system it would try to represent itself as a nation, and argue that its influence would need to be just as great as a similar sized nation. After all if the AI would represent just a single vote it would be vulnerable to politics of other nations/political groups in the GD. The AI would never join if at minimum it would not have the same political oomph as any nation or group of similar size, so that would be the minimal accepted. Any alterations would need to come from your idea of the political structure and cultures governing the GD.

$\endgroup$
3
  • $\begingroup$ If the government recognises the newly made small AIs as citizens then it’s not cheating. It’s just very rapid population growth. $\endgroup$
    – Joe Bloggs
    Commented Jul 11, 2021 at 8:41
  • 3
    $\begingroup$ @JoeBloggs if you store a few trillion small AI's somewhere for votes you can essentially build political power. I doubt such a move would be accepted because of its potential abuse. $\endgroup$
    – Demigan
    Commented Jul 11, 2021 at 11:43
  • 2
    $\begingroup$ @JoeBloggs Any other race can probably manufacture citizens as well -- say the space-frogs lay 10,000 tadpoles and only 2 survive to juvenile frogs. I think your example shows 1 person/1 galactic vote in galactic congress isn't really workable. $\endgroup$ Commented Jul 11, 2021 at 16:13
4
$\begingroup$

The AI is a voting block, not an individual

The AI has the combined output of the whole solar system, meaning in terms of what it can do and its value it is the same as a group of people who inhabit the solar system and do what it does. While it might not make sense to give it individual votes for a popular vote it makes sense to give it electoral votes, senators, and representatives. But how many votes should it get?

No taxation without representation

The reason we have representatives and senators is to balance the power of small states and large states. However, ultimately every state stays in the union because the political power it derives from representation and services from being in the union outweighs the costs of the being in the union. If a state lost some of its political power in terms of representatives then it might cause them to leave since their citizens' lack of agency in politics leaves them vulnerable. Therefore, the amount of votes the AI gets is equal to the minimum amount of votes you can give it without it threatening to withdraw from the Galactic democracy.

Start by taking a solar system of a similar size and give the AI that many votes. Anything less and the AI might reconsider joining.

$\endgroup$
3
$\begingroup$

Political compromise based on proxies

The only workable compromise in the long run is one which gives the AI-controlled system political influence roughly in proportion to its contribution to the federation (so no grossly lopsided scenarios where the AI is limited to one vote, or can game the system by creating infinite bot-citizens). I suggest a very rough baseline for the votes it receives would be to use economic production as a proxy for population i.e. take the Gross National Product (or Gross System Product) of the AI, then divide it by the federation's average GNP per capita to generate a notional population figure for the AI's system. Further adjust it for the federation's average rate of enfranchisement (i.e. if 30% of the galactic population cannot vote because they are children/larvae/clone-serfs, knock 30% off the AI's vote total too), and give the AI that many votes.

Of course, this is a very crude approximation metric, and the final figure will be subject to political haggling based on a range of specific arguments (how directly comparable are different kinds of production in your futuristic economy? Is there another metric alongside economic production which better represents the federation's notion of value? How does the law treat lesser sentient AIs that already exist within federation member-states' territory? Etc etc). In particular, I expect federation negotiators will want to set the AI's vote total slightly lower than the raw population-proxy numbers will suggest, because it has a number of advantages which could otherwise give it disproportionate power. The AI is a single superintelligent consciousness with complete control over all its votes, so it can achieve a 100% turnout rate (which other systems presumably cannot) and tactically direct every single ballot to achieve maximum impact; this will need to be offset somehow.

The final compromise that emerges will thus depend on a vast range of political variables that I can't estimate based on the information provided, but I've suggested one plausible framework which could guide accession negotiations. The wider functioning of a galactic democracy will be built on such compromises, because 'one person, one vote' is tricky when future-tech and radical inter-species differences make 'person' a fluid and contested concept.

$\endgroup$
4
  • $\begingroup$ "roughly in proportion to its contribution to the federation" that was good actually, but the you dived sidewise, imh. Was a good premise // ahh sorry forgot ops flawed pemise, then it's fine answer $\endgroup$
    – MolbOrg
    Commented Jul 11, 2021 at 18:53
  • $\begingroup$ I wonder if economics are the right way to do this. A race with billions of smaller low intelligence creatures might have so little economic output that it requires support, but would those billions really have less voting capacity as a race of powerful high intelligence beings that produce tons more per capita? It would make it less a democracy and more an Oligarchy as the richest and best producing races will get the most votes. This also depends on the resources in the solar systems they live in, the length of the trade routes etc. $\endgroup$
    – Demigan
    Commented Jul 11, 2021 at 19:04
  • $\begingroup$ @Demigan I agree that economic output shouldn't be used as a general means of assigning votes across the federation; that way plutocracy lies. But it is one potential way of dealing with the unique problem of the AI, a system-spanning civilisation which can't be divided into discrete individuals. My answer suggested that GNP might also be only one part of an aggregate measure of value, which could perhaps incorporate metrics for cultural/scientific/military/other contributions. Perhaps you could have a need-based metric too (e.g. the AI's power-consumption requirements)? $\endgroup$ Commented Jul 12, 2021 at 8:57
  • $\begingroup$ @Demigan not economics, or income per capita - contribution to galaxy thing - u give more resources for common cause, weigth of your voting is more in descisions where those resources go, how they are used and if one contributes with something else there is some corresponding resource value as well. $\endgroup$
    – MolbOrg
    Commented Jul 12, 2021 at 17:16
2
$\begingroup$

No solution is perfect, but I'd argue in favor of one vote.

Here's the thing-- one vote can both represent the needs of an individual and state. On issues of morality or opinion, one vote is really all a single intelligent being should expect, regardless of power and position.

On the other hand, in issues regarding member states, having only one or two votes would probably be the norm anyway. Kind of like how the U.S. Senate only has two representatives of each state.

So yes, in general elections, the AI only deserves one vote (as one being); in legislative votes, the AI should get the same as every other member state receives (likely one vote).

$\endgroup$
2
$\begingroup$

None!

"How many votes" depends on what it gets to vote on. Analysis of that question reveals that all answers are absurd.

Consider the real-world U.S.: rank-and-file voters get to vote for people they send to the national Congress, but they don't get to vote on e.g. the For the People Act, or the John Lewis Voting Rights Act, or whether to hear witnesses and see evidence in either of Donald Trump's impeachment trials. Generally, they vote for representatives, not on specific questions faced by the government.

It is absurd to think that a massive AI that spans a solar system would vote for a representative of any kind: some crude meat puppet whose job it is to "speak on behalf of" the what is probably most intelligent entity in existence.

The AI would itself be its own Senators, Representatives -- permanently and not subject to the votes or wishes of any other political region within the democracy. And, if there is a President, it seems like the AI would need to be eligible to run for that office.

$\endgroup$
5
  • 1
    $\begingroup$ So, how many Senators and Representatives would the AI get to play then? $\endgroup$
    – ilkkachu
    Commented Jul 11, 2021 at 20:17
  • $\begingroup$ @ilkkachu It would depend on how representation is allocated. OP didn't share those details, so it's impossible to say. If it's based on territory, like the real U.S. Senate, then it would be identical to a scenario in which there is no AI and the system is populated by organic life. If it's based on something else... well, it'll be something else, which may translate more or less cleanly onto an AI scenario. $\endgroup$
    – Tom
    Commented Jul 11, 2021 at 21:39
  • 2
    $\begingroup$ I think the question was about something population-based, i.e. an analogon to the House of Representatives in the US. $\endgroup$ Commented Jul 11, 2021 at 23:17
  • $\begingroup$ @PaŭloEbermann I have a hunch you're right, but I'd ask OP to state that explicitly, not just to eliminate ambiguity, but to nail down the details so that we can see what is and isn't nailed down, and so we can try to reasonably extrapolate from OP's established rules. I don't think it's safe for everyone on WB.SE to assume that the United States' peculiar institutions are the starting point of every question. $\endgroup$
    – Tom
    Commented Jul 12, 2021 at 0:33
  • $\begingroup$ @Tom I think you're probably right that it's easier to directly assign the AI seats in the Galactic Parliament rather than votes, but this likely still requires some notional figure for its 'electorate'; it's not a US peculiarity to assign seats in at least one chamber according to population, all but the loosest federal democracies (e.g. the UN) do this. And if there are direct elections for galaxy-wide offices (e.g. president) we need to know how many votes the AI gets. So the core problem is probably unchanged, though I agree that more info on the federation's constitution would be useful. $\endgroup$ Commented Jul 13, 2021 at 11:04
2
$\begingroup$

a multispecies Galactic Democracy with Proportional Representation

If the Galactic Democracy is to keep its status as a credible political entity, they should probably give the AI-state votes by the same rules they've given them to everyone else.

And especially since you specified it's a multispecies coalition, that probably isn't going to be one vote per individual, for multiple reasons. The member species are likely to be different, in their history and culture, governance within their planet/system/whatever but also in their biology. One species might be biologically smaller than another, and be able to support a greater population on a planet of a similar size and economic output. (Well, unless all your species are rubber-masked humanoids, of course.)

Regardless of biology, some planet might be more densely populated, while another is sparser. (Slaves might be a problem too, unless the GD has managed to outlaw that totally.) It's probably not a good idea for the GD to encourage the member species to overpopulate their planets to get more political oomph either, if it cares about the well-being of the individuals living down there.

Those kinds of things would make the "one biological entity, one vote" system problematic even before the AI comes into play, and it's far more likely that the GD would be set up more as a Galactic Federation, with some sort of representative democracy at the top level, with varying levels of voting power based on economic or political strength or such.

Note that, even just with humans in play, in the United Nations General Assembly, each country has just one vote, not one vote per citizen/inhabitant. (And then there's the Security Council.)

$\endgroup$
1
$\begingroup$

This is a difficult question because it is very much opinion-based, with no clear answer. There is already one close vote as I write this. However, I do not vote to close myself because there are some elements of an answer which may be relevant:

  • Presumably your galactic civilization has different, sentient species. How do they deal with different ages of majority/consent across species lines? There could be a species where only one gender, or only the very elderly, are allowed to vote.
    One option would be to have a standardized test that is applied to representatives of the species to determine reasoning ability. "At the age of 14.7 standard years, the 4th quintile of the sample population has reached the reasoning standards. Therefore members of this species are deemed adult at 14.7 standard years unless the individual has been judged incompetent by the legal authorities."
    How many subprocesses can the AI spawn at the same time that would pass the test? Each gets one vote.
  • Or possibly there are precedents for a hive mind in the confederation. Are drones allowed to vote if they have no free will?
$\endgroup$
1
$\begingroup$

Taxation for service.

They would be taxed based on the number of robots they can support with their intellect, with a necessity to have a certain number of a certain type of robots available for use.

Intelligent, well managed people are useful. The galactic nation has been able to call up it's voting population many times for war, producing large amounts of some valued good, or soldiers to crush an evil invader. They would negotiate for a while and come to a fair compromise over how many votes would be granted for each bit of computing power.

No doubt other nations have come upon the strategy of producing lots of people to get more votes. This AI system can do the same, but then it needs to be ready to contribute more soldiers or production robots and metal minds to necessary galaxy wide war efforts.

$\endgroup$
1
$\begingroup$

GD as Corporation

Make the members of the GD purchase their votes via shares. This is a practical way to implement Montefeltro's observation that political power should be proportional to a member's "contribution", which is generally understood to be related to their economic output. It also allows some systems to effectively opt out of participation, and allows other more enthusiastic systems to gain more than they otherwise would.

IPO

Initially, all votes would need to be distributed to all systems in some person-agnostic manner. I would propose this be done on the basis of the blackbody temperature of the system. This measures the amount of energy available to the system, and indirectly, its mass. If the GD contains Kardashev Type I/II civilizations, then trying to measure anything less than this is futile and useless. If all the civs are < Type I, then I question how a GD is even possible, since they probably only have slow interstellar travel.

One the shares are allocated, they can then be traded on the open market. Of course, there should be more than one share per system during the IPO. A target should be something like 1,000 or 1,000,000 shares per system, average. This would allow members to adjust their share pool up or down by buying and selling a fraction of their representation (based partly on the level of their system exploitation...a a system with an energetic star but poorly habitable planets may decide that a more powerful civilization could better utilize its resources, and choose to sell many of its shares for a big economic bump, rather than trying to wield them all in galactic matters of only minor import to them).

Monopoly

There is, of course, the incentive for one or a small number of members to hoard votes. But who would they hoard them from? They have to buy every vote from someone else, and the more they amass, the more the rest of the GD will bid up votes, just like in any market. It would be irrational for any star system to charge less than the total economic output of the galaxy to purchase the last share, or even the share which achieves majority, whatever threshold that is. Every star system will be eyeing every other which is buying up shares, trying to determine their motives, their ability, their aggression, etc. If any system is able to corner a significant portion of the market, then it will only be because the majority voluntarily sold their shares. This means they either trust the majoritarian, or they are so economically/militarily/politically weak that they would have been steamrolled in any other system of gov't as well.

Power

All of this discussion is moot without knowing what individual members are capable of w.r.t. power projection across the galaxy. Is there FTL travel? FTL signalling? How is trade conducted? What goods do members consider valuable enough to trade between stars? There must be trade, or there is no incentive for the most powerful system to simply conquer all of its weaker rivals. Unless, of course, the weaker systems are collectively powerful enough to resist the powerful one. But holding such a coalition together without trade relies 100% on game theory, and is likely very unstable. If even a few of the members decide that they are better off siding with the big dog, then the coalition falls apart.

What are the military capabilities of the member states? Can they destroy planets? Blockade them? How can they interfere with trade, given that space is so big? Can they kill all the citizens on a planet while leaving the infrastructure? Are the species remotely biologically compatible, or are they a hodgepodge mix? Are they at least all carbon-based, with some organic compounds in common?

$\endgroup$
1
  • $\begingroup$ Okay, some good answer - do not forget renting votes, for some descisions, or time - so small entities have a chance to focus on matters which are important for them. It also may be a total replacemnt for selig, prevent hording, and means to disrupt monopolies if they decide cash out reputation and good stuff. $\endgroup$
    – MolbOrg
    Commented Jul 11, 2021 at 19:10
0
$\begingroup$

One AI, one vote.

Otherwise you are arguing that the rich and powerful are entitled to more votes than the poor and weak. That can of worms would result in probable war and certain civil disorder.

$\endgroup$
0
$\begingroup$

Option A one vote

I think the fundamental question here - actually preceeding the opening - is whether these "democratic galactic values" would include acceptance of AI beings as aware individuals with certain rights. It should meet certain criteria for that.. would the AI be able to shape (and refine) its own value system ? Develop its own opinions, independent of its designers ? If that would be the case, the AI system will be entitled to ONE vote, just because it is being acknowledged as an adult individual.

Passive suffrage too

When the AI would be assigned all democratic rights, it could also become a candidate people vote for. Suppose the AI has always been a peaceful and succesful entity "controlling the solar system", chance is, people in its solar system would be prepared to vote for it, in large numbers !

$\endgroup$
1
  • $\begingroup$ If I understand the question right, the premise is that there are no (biological) people in the solar system governed by that AI. $\endgroup$ Commented Jul 11, 2021 at 23:19
0
$\begingroup$

Being a single AI, it can be expected to vote systematically to get as much influence from its voting power as possible. If you want your democracy to remain stable, you should not give it too much power, obviously; Otherwise, the democracy might quickly end up with some Caesar/Hitler/Palpatine too strong to be removed by democratic means, effectively turning the democracy into an AI-cracy. Up to several % of the total voting pool should probably be safe, unless the AI can build that up by collaborating with other similar entities.

Inside the stability limit, I think the only criterion is to keep some vague sense of fairness, so the inhabitants do not tend to start a revolution. Be it based on taxes, GDP, military service (Starship Troopers), or something arbitrary (USA). Maybe it's not such a bad idea to give the AI just one vote, so nobody else feels offended. The AI may agree to that since it realizes there are other ways of power in a democracy (like becoming a politician, or just contributing to economy, if we stick to completely legal ways). If not, try something else from the safe side. Make it one "elector", as in US presidential elections? I don't think it really matters (better do not tell that to voters).

That being said, AI citizens will definitely need some legal protection, fundamental rights, etc. (unless their military or economical power is just big enough...) Conspiracies will go both ways. People should not be able to decide that all AIs are slaves or something. Just voting against the anti-AI party may not be enough here.

$\endgroup$

Not the answer you're looking for? Browse other questions tagged .