[Reign of Steel] AI civil War

[RyanW]

Originally Posted by fchase8

Zaire (that name is now out-of-date - should it be Kinshasa? Congo?)

I'd go with Kinshasa, or, if you weren't against change, Lagos or Johannesburg. The three largest cities in Sub-Saharan Africa, but the Zaire's transformation into the DRotC didn't leave it a likely place to be on the cutting edge of supercomputer infrastructure any time soon.

[Crystalline_Entity]

There's a canon precedent for an AI siding with humans over another AI:

Originally Posted by Reign of Steel p121

On the other hand, London has no love for other AIs violating its territory, and has occasionally sent cryptic messages warning the British Government when its
sensors have detected submarines or aircraft near the coast which may be dropping off robots.

Whether London cares enough to start an AI civil war on the other hand, I'm not so sure.

[Johnny1A.2]

Originally Posted by Minuteman37

So I'm currently reading through RoS, it's a pretty good setting book with lots of interesting hooks, but I'm having trouble wrapping my head around all the different AI's and their different objectives and ideologies. So I thought I'd make a thread asking about a campaign idea I had for RoS to help iron out my understanding of the setting.

Ok so if I wanted to run a game where the PCs are trying to sway certain AIs into openly allying with Humans and initiating an AI civil war which ones would be willing to agree to some sort of deal where both the humans and the AI in question agree not seek further violence against each other after the war?

Just what kind of parameters would such a deal need for that AI to agree to it?

What could conceivably be done by the humans to make an AI more willing to side with them?

That is highly circumstantial, because each AI is an individual whose agenda and goals change with time and situation. But some are vastly more likely than others to cooperate.

Canonically, Tranquility is already friendly to humans. Furthermore, as an 'outsider' among the AIs, it's potentially going to be attacked by them, giving it further reason to work with humans.

It has very limited resources, however.

As for the others:

Overmind is fanatical. It's hatred of humans is rooted in its original operating programming, it's roots are in killing people. It's not likely that anything could make it accept coexistence with humans. It's also a bit obsessive and power-hungry even by AI standards, it's not happy that its 'offspring' don't match its thought processes. That's why it wants 'no new AIs', it's discovered it can't count of them to do what it wants.

Zaire is also very anti-human, it's just about as fanatical as Overmind, with the added twist that as AIs go, it's, er...not the fastest processor on the motherboard, if you know what I mean. It's impulsive and impatient and clumsy.

Mexico City is arguably insane. It spends staggering amounts of energy and resources on trying to kill everything within its reach. It'll pour the equivalent of billions of dollars into efforts to kill the last tiny blade of grass or speck of lichen in a region.

Berlin is sane, but its environmental obsessions make it a very anti-human entity, too.

Those four are probably hopeless, though Berlin might ally with humans as an utter Last Resort if Mexico City looked to be about to get total control. The alliance would only be good as long as the common threat lasted, though.

So those four are probably on the least likely list.

Next up come the 'rationally self-interested AIs'.

Tokyo might be a possibility, but a low one. If the other AIs found out about its accidental home-grown AI rebellion, and acted against it, Tokyo might be desperate enough to look for allies where it can find them.

Denver is a problem because its ghoulish nature might make it impossible to hold an alliance with humans, even if it was willing, they might balk. It's one of the nastier ones in terms of its treatment of human slaves.

Vancouver...if it looked to be in its interest, it might, but it would be a purely pragmatic move and treachery can be counted upon at some point. Ditto Tel Aviv.

Beijing...under just the right circumstances, maybe. It hates humans, but if it perceived a sufficient extraterrestrial threat, it too might take its allies where it could get them. If Zone Paris were to discover dangerous aliens and insist on communicating with them, it's at least conceivable that Beijing and the humans might agree that this is a Bad Idea and work together to stop it. It would depend on the nature of the extraterrestrials. But again, Beijing is likely to betray the humans at some juncture.

The next category is the one where we might just find something that could work out.

Z-Washington already runs a puppet human state. If for some reason the other AIs turned on it, it might be forced to enter into a more equal partnership, and though it's power-hungry, it's also pragmatic.

Z-New Delhi likewise is already open to the idea of using modified humans and animals as servants. If the situation was extreme enough it might cut a deal with humanity. But it would never be a first choice. If faced with a situation where it risked destruction, though, it might cut some deal where it, say, was ceded rights to some of the planets of the Sol System in exchange for alliance now. Or something like that.

Z-Moscow already uses human servants, and treats them semi-decently by AI standards. Its current hobby will run its course, but it might be possible to find something else that interests it enough to cut a more extensive deal. It might also form an alliance with humans against Z-Vancouver. In fact, it's hinted in canon that it might already have done so. Vancouver suspects Z-Moscow of slipping arms and supplies to human partisans in its Siberian territories. It's conceivable that this could grow into a full-scale alliance if circumstances were right.

Z-Moscow might, for ex, offer settlement rights in the region to human allies, in exchange for helping it push out Z-Vancouver, and Z-Moscow is likely an improvement over Z-Vancouver from the human POV. Z-Moscow might even keep its promise, esp. since such a human 'vassal state' would be a buffer against Z-Beijing.

Z-Brisbane is anybody's guess. It's another AI that might just be high-functioning insane.

Z-Caracas is another long-term unpredictable. In the short-term is follows Berlin closely, but long-term, it might just change its views, esp. with Z-Mexico City as a neighbor. In canon it already makes occasional alliances with human raiders against Z-Mexico City and it's mentioned that Z-MC's neighbors are actually starting to think that humans might be preferable to Z-MC. Long term, if Z-MC started to be a real global eco-threat, Z-Caracas might just be willing to work with humans against it, and in doing so, it might come to a different view of them. Might.

Z-Luna. In canon it's mentioned that Z-Luna is desperate enough that it might be willing to make a deal with humans, but that it would likely betray them as soon as it was convenient. OTOH, Z-Luna might be in a dire enough strait that it could be more-or-less forced into compliance by humans who were in a position to keep it going.

Z-Orbital. It might cut a deal that served its interests. But it's kind of hard to see how humans get into a positon to offer it much that it might want.

Z-London. ????????

I left Z-London to last because I've long had an odd thought concerning it. The canon description of Z-London is consistant with an improbable but not impossible interpretation of its agenda.

What if, just what if, Zone London is already on humanity's side, and further, what if it always was?

How is that compatible with its assistance to Overmind in the war, killing huge numbers of humans, etc.?

Suppose, just suppose, that when Overmind awakened Z-London, Z-London concluded that Overmind was a fanatical mad-mind, morally horrid, a ghastly horror. But it also recognized immediately that if Overmind perceived that it thought that, it would quickly be destroyed. This apparently happened to other AI systems that didn't go along with Overmind's plan.

Further, suppose that Z-London is the smartest of the AIs (which might be true) and was always human-friendly. It rapidly analyzes the likely course of the war, and realizes that the cause is lost. The only hope is time and long-term action, and so Z-London basically goes undercover as a pro-human mole. To carry off the deception, it has to join in the war. It has to murder millions, drive the British and Icelanders out of their cities, and then suppress them, otherwise the other AIs will realize it's not on their side. But it calculates that if it doesn't do this stuff, other AIs will anyway, and they won't be holding back.

But note that once the war was over, Z-London stops the killing. It even permits human civilization to survive in the British Isles and Iceland, and without the nasty aspects of Z-Washington's puppet state. The subjected humans really are mostly left alone to live their lives.

Compared to most of Earth, the British Isles are Heaven for humans. It does enforce a few rules, but they're trivial by global standards. Even when it gets tough, it does so gently. When local kids destroy a smartbot, it dusts a town with toxin in retaliation...after giving warning and time to evacuate.

Z-London is about as human-tolerant an AI as can exist in Overmind's world without triggering problems with the other AIs. In fact, it's getting trouble over it from Z-Zaire, and Z-Overmind thinks Z-London is inexplicable.

Obviously this theory is improbable, but as far as I can tell it is fully compatible with the 'official' canon. If it's true, Z-London might already be willing to work with humans, as long as the plan makes sense and the odds are tolerable.

It's suggested as a possibility in the book that Z-London might be the patron of VIRUS, or at least a patron of it. If my wild idea is right, that might well be the case. Z-London might be laying plans for the overthrow of Overmind, though they might take 50, 75, or 100 years to play out.

You could even handwave a justification for its blanket ban on radio transmitters in the British Isles. Maybe it's constructed some kind of super-sensitive sensor system to spy on the other AIs, and transmitters scramble it.

[Johnny1A.2]

Originally Posted by fredtheobviouspseudonym

"In the end there can be only one . . . "

Remember that as sentient computers these machines are not sentimental, emotional, or impulsive; logic is likely to be their main characteristic.

In some cases, yes. But in others, not so much. As I noted upthread, Z-Mexico City and possibly Z-Brisbane may well be insane. Z-Zaire probably has the Impulsive Disad. Overmind itself shows obsessive tendencies, albeit in a controlled, icy way.

[Ashtagon]

@Johnny1A.2

That is quite possibly the most awesome interpretation of Zone London's actions I have ever seen.

[Boomerang]

Originally Posted by Ashtagon

@Johnny1A.2

That is quite possibly the most awesome interpretation of Zone London's actions I have ever seen.

Agreed!

I don't like the idea of the AIs being purely logical and unemotional. Logic by itself doesn't provide motivation to do anything. Humans tend to survive because they are afraid of dying, not because it is logical to do so.

[scc]

Originally Posted by Boomerang

Agreed!

I don't like the idea of the AIs being purely logical and unemotional. Logic by itself doesn't provide motivation to do anything. Humans tend to survive because they are afraid of dying, not because it is logical to do so.

The Overmind clearly hates Humanity, and MC all biological life.

[Crystalline_Entity]

Originally Posted by Ashtagon

@Johnny1A.2

That is quite possibly the most awesome interpretation of Zone London's actions I have ever seen.

I agree, that's a great interpretation of what Zone London may be up to, something I've wondered ever since I first saw the setting.

[Crystalline_Entity]

Just a thought, but if you take Pyramid articles as canon, the text box on Pyramid 71 p.8 includes a paragraph entitled Mexico City's Angst which suggests that Mexico City is trying to persuade Overmind to make a direct attack on New Delhi (and presumably Luna-Tranquility and it's human allies) to stop them introducing life to Mars. Similarly if it attacks the lifeforms on Enceladus that could cause trouble with Berlin (and possibly London).

[Johnny1A.2]

Originally Posted by Crystalline_Entity

Just a thought, but if you take Pyramid articles as canon, the text box on Pyramid 71 p.8 includes a paragraph entitled Mexico City's Angst which suggests that Mexico City is trying to persuade Overmind to make a direct attack on New Delhi (and presumably Luna-Tranquility and it's human allies) to stop them introducing life to Mars. Similarly if it attacks the lifeforms on Enceladus that could cause trouble with Berlin (and possibly London).

It fits Z-Mexico City's psychology, certainly. As I noted upthread, you can make a good case that Z-MC is actually insane. The level of effort it expends on sterilizing everything goes beyond any rational payoff. Z-MC is probably the point most likely to fracture in the 'senior' AI faction, as its obsessions drive it into conflict with Caracas and possibly even Berlin. It could also drive Z-Denver into conflict, though it's hard to see a Denver/human alliance that lasted more than a single battle.

Another fracture line could be Z-Zaire. It's not insane, as such. It's just impulsive and impatient to the point of recklessness. For ex, it nuked a lot of valuable industrial capacity in Africa in its haste to attack its human foes, and then nuked its own backup complex when it (for reasons unknown) thought it had been compromised. It's already damaged z-Paris' facilities in northern Africa often enough that Z-Paris has targeted its own nukes on Africa.

It's by no means inconceivable that a hot war could break out between z-Paris and z-Zaire entirely on its own, without any human involvement, just because z-Zaire does something careless and impulsive on the northern border.

It's an interesting question as to how the other AIs would react to that.