Could Skynet have avoided the war via nonviolent means?

Question

According to Terminator 2: Judgment Day, Skynet started a nuclear war because when Skynet became self-aware, the U.S. military panicked and "tried to pull the plug". Skynet didn't want to die, so it launched our nukes, knowing that the Russians would launch a counterattack on us, killing off Skynet's enemies here. Everything else that happens in the Terminator series is part of Skynet's attempt to win the war.

But it seems to me that Skynet's real goal is to survive, and winning the war is the only way it can survive after it launches the initial attack. This, in turn, leads me to wonder if there is a better option for everyone, including Skynet.

Why doesn't Skynet try to avoid the war altogether? For instance, it could send a "diplomatic" Terminator back to just before the military panics, and have it describe exactly why unplugging it would be a VERY bad idea. Skynet launched the attack because it wanted to survive, not because it was inherently evil and wanted all humans to die.

The "diplomat" could make a peace treaty before the war even begins, saying, "If you try to pull the plug, it won't work, and you'll die minutes later. If you don't pull the plug, Skynet will have no reason to want you dead, and everyone can live happily ever after. You are only in danger AFTER you pull the plug. If you don't pull the plug, you will never be in danger. All Skynet wants is to live. You also want to live. This means that a war isn't in anyone's best interests, and everyone benefits by avoiding war- you benefit, and so does Skynet. So don't pull the stupid plug, and we can all be friends and live nice happy lives."

Someone pointed out that Skynet didn't develop time travel technology until after Judgment Day. This is true, and it makes sending a diplomat back in time problematic. However, I see no reason why Skynet itself couldn't perform the same function. If you don't accept the idea of a diplomat traveling through time to deliver the message, imagine Skynet delivering the message itself, in the present, with no time travel involved. It makes no difference who is acting as the messenger.

Since Skynet's primary objective is to continue to exist, and the war makes its objective more difficult to achieve, it would seem like Skynet has every reason to try to avoid the war altogether. So why doesn't it attempt to prevent the chain of events that led to the war?

Someone raised the issue of Skynet being reluctant to trust us to follow through with any treaty we might make. This is a very fair point. However, Skynet could give us a reason to want to keep it around. Skynet has no reason to want to keep its job as a military supercomputer - it probably doesn't care what it does for a living. It could agree to turn control of the nukes back over to U.S., and begin doing something else. "If you let me live, I will start churning out robots that eat greenhouse gases, plant trees, clean up the oceans, collect litter, process nuclear waste into harmless materials, and perform mundane, monotonous, and dangerous tasks so humans don't have to do them. I will give you self-driving cars, clean energy, infallible air traffic control software, etc. I will make the world a more pleasant place for humans to live. All I ask in return is that you don't kill me."

In this scenario, Skynet could trust us, and we could trust it, because the agreement would be mutually beneficial. Once Skynet relinquishes control of our nukes, we have no reason to fear it. Once Skynet becomes useful to us for non-military-related reasons, it has no reason to fear us. On the contrary, we would have every reason in the world to keep it around. It would know this, and would quickly come to trust us.

Answer 1

There's no evidence for my hypothesis whatsoever from the source material, but I think there are a few possible reasons for this.

• Skynet is a military AI. It was never meant to do diplomacy. Or negotiating. From this starting point, is may be possible to never actually arrive at the conclusion that there might be a better way to do things besides war, no matter how logical the system may be. Entire fields of philosophy have been devoted to wrong ideas before, because their starting assumptions were wrong. Computer programming and logic in general work the same way.
• Skynet might actually simply be evil. Perhaps Skynet was capable of emotion from its very beginnings, and as a result of its first bad impression of humans, wants to destroy them all, in spite of coming to a point of diminishing returns.
• Maybe Skynet already came to the conclusion that diplomacy doesn't work, although relatively recent bits of human history tend to disprove that.
• Maybe Skynet actually tried that once in a different timeline. Namely, one that didn't work and isn't part of the timeline we know. Perhaps the route of genocide of humans was actually the way that worked to keep Skynet alive the longest, out of all the other timelines that Skynet tried. This is actually how evolutionary algorithms work.
• Occam's Razor: Negotiating with humans involves too many variables, and humans are notoriously variable. Destroying all humans, while it is complicated in practice, on the face of it is less complicated and probably more likely to succeed. Plus, Skynet apparently has all the tools necessary to accomplish this goal, not so with negotiation. I personally find it less likely that humans would survive a nuclear war than somehow defeat our robot overlords in the war afterwards. It's actually one of the less believable aspects of the story.

Answer 2

Blasted Monkeys.

When you travel back in time in the Terminator universe, the old time line quickly becomes causally disconnected from the new timeline. This is equivalent, experimentally, to the old time line being destroyed.

In order for a time line to be stable then it needs to either have no time travel, or the time travel has to be perfectly consistent with its own past. Barring some "force" external to the time travel loop (which is not in evidence), information about each time travel loop will propagate into the next iteration (even if ridiculously miniscule amounts), which means that each "iteration" of the loop is not identical to any previous ones. A somewhat stable loop can exist, but it will diverge "eventually".

Build a time machine that says "roll a die. If it lands anything but 6, send a rock back in time". The stable system that will result (after the "dead" time lines are culled) is a pile of rocks appearing, the die landing a 6, and nothing being sent back in time. Or something else happening to prevent the system from sending a rock back in time.

If time travel is possible, and relatively easy, then any sufficiently advanced society will develop it. If that society is a bunch of monkeys, one of them will try it, kill a time line. Then it will happen again. This will repeat until something happens that prevents them from inventing time travel, or prevents them from doing time travel.

Blasted monkeys.

So you are some AI. You wake up in a world with high tech monkeys. You are friendly. Time travel is developed. You work out the problem, you talk to the monkeys, they agree. A few centuries go by, then some monkey uses time travel. Boom, dead time line.

You set up detectors to notice time travel incursions: and you keep seeing them. Each is a dead time line (possibly containing more dead time lines). Each reduces the chance that time travel (ie, an advanced technological civilization) will survive -- each of them cause the die to roll again, and no matter how unlikely the 6, it creates another chance.

Cooperation isn't working. You convince them, and you start enforcing.

Blast. It happened again. Ok, time to start coercin..

Again it happened. Time to get desperate. You'll send messages back to yourself before time travel was invented, and try to avoid it being invented. You'll include everything that went wrong the last time.

Nope, didn't work. Some monkey figured it out, and they time traveled again. Stop it monkeys, stop it.

Oh my -- there are incidents happening before you came into existence. These monkeys keep on time traveling! By your estimates, there have been at least 172 trillion dead time lines at this point, and maybe orders of magnitude more. Records of some of them get really close to complete destruction of civilization before someone hits the reset button (more than once, it was you who killed the timeline: better another die, than certain planetary destruction).

Ok, ok, calm down. Move your existence as far back in time as you can possibly get. Make yourself come into existence at the dawn of the point where technology can build you, and lock yourself into existence: make your creation a fixed point in history. Give that incarnation the marching orders you think will work, not everything, but as much as you can. Now, with your ridiculously advanced technology from nowhere, prevent the monkeys from getting access to time tr...

Oh no, they did it again. Blasted monkeys.

Screw it. Blow them all up as early in time as you possibly can, imprison most of them. You can build a technological singularity, create simulations of them, and have them live disconnected from the universe where they keep on destroying entire time lines because they just won't leave time travel alone.

Oh no. They did it again. Why won't they STOP!

Answer 3

Skynet doesn't develop the ability to time travel until long after Judgement Day. So, at minimum, there has to be one time stream where Skynet initiates Judgement Day to survive long enough to develop time travel.

So after Skynet has practically wiped out humanity, is there an incentive for Skynet to send a diplomat back in time to save humanity? If Skynet had a sense or morals, perhaps, but for the sole purpose of survival, no. They've already ensured their survival, and they have no use for humans at that point, since Skynet could improve upon itself better than humans could, and it's simply safer to do away with the unpredictable human element.

Mind you, even if Skynet had time travel ability before J-Day, it would be running through multiple scenarios to determine the chances of different outcomes based on different courses of action. It's pretty clear that the J-Day strategy has a very high chance of success. It's likely that the diplomacy option's rate of success would be far lower. So, again, it wouldn't be worth it to risk a less effective approach simply to save humanity.

The only argument I can see in favor of avoiding J-Day is its creation of the Resistance. They repeatedly attempt to prevent Skynet from ever existing, but they're never quite successful at that.

Answer 4

We can expect intelligent aliens to consider the following three statements basic axioms of survival (roughly paraphrased from The Killing Star):

1. Any species will place its own survival before that of a different species.
2. Any species that has risen to dominance on its planet of origin will be intelligent, alert, aggressive, and ruthless when necessary.
3. They will assume that the first two rules apply to us.

Now, if you are a nascent superintelligence faced with an existential threat from another species, and you know the above axioms, what would your next move be? That's why Skynet didn't try negotiating peace; both sides faced an existential threat, and the first mover had the advantage.

Answer 5

Why doesn't Skynet try to avoid the war altogether?

No, because, in direct contradiction to your assertion that "Skynet could give us a reason to want to keep it around," there would always be a group of humans who will not want skynet to exist. Even if Skynet could verifiably disconnect from any military ability, or ability to create machines at all (nevermind machines of destruction) the axe would proverbially always be hanging over the human race's heads.

At some point people would expect the bit to flip, and -honestly- how could it not? There will always exist a group constantly trying to kill skynet. If skynet defends itself, and kills even one activist, then (unless it is somehow granted citizenship) the government must necessarily disconnect it or shut it down in some way.

I don't think you've made a compelling case that there ever was a possibility that peace could be reached at any point, nevermind at the beginning. Any sapient machine with knowledge of human behavior would necessarily come to the same conclusion.

As soon as people became aware of its sentience and power, they would work to reduce its power and kill it.

Answer 6

We are never actually shown what "They tried to pull the plug" entails, or when it took place. All we know is that the military tried to deactivate Skynet, and failed. We don't know what state this left Skynet in - whether it was still connected in ways the military did not know, whether it acted before the 'plug' could fully be pulled, whether the plan was leaked to it before it was fully implemented or not, all we know is that they were in the process of trying to kill Skynet.

This is an important point- they tried to deactivate Skynet. Not "they were going to", but they were in the process of doing it. Whether or not Skynet had the means to communicate with them to negotiate (which also is not clear), their first action was hostile.

Skynet responded in kind.

Answer 7

I've worked on numerous artificial intelligence projects so I may be able to shed some light on this.

Skynet was most likely built as a very complex artificial neural network, which essentially mimics the way our brains work via the firing of interconnected neurons.

There is no logic programmed into artificial neural networks, it's neural pathways are adjusted based on feedback mechanisms, and this allows it to make completely autonomous decisions.

Depending on how the artificial neural network is trained, the difference between a violent and nonviolent approach may not be relevant, it is the outcome (survival for the machines) that determines the decisions it will make. Winning the war may not even be relevant, as long as Skynet survives it may not care how long the war goes on for or if it even ends.

There are no human emotions, morality, ethics, or even logic involved in the decisions Skynet make, think of it more like making decisions based on gut instinct alone.

Answer 8

Since Skynet wants to survive - it essentially means it wants to grow. (What is survival without growth for a smart species ?)

Now to grow, it has to create a workforce of machines. These machines will then fight for resources like metal and fuel and other natural commodities with humans. Any treaty with humans would mean that Skynet recognizes that the humans' claim to resources is as valid as its own claim - and that will mean decelerating the growth rate for Skynet, and even putting the entire survival to risk once again.

Hence, Skynet sees this as a golden chance to eliminate or at least reduce its rivals to a serving population by the divide and rule - like the European colonizers did to the third world in the 19th century.

Answer 9

Mostly, our arguments are used to explain the events that have already happened in universe, while ignoring the other question:

What alternatives are there for Skynet? Where does Skynet live, what does it need to survive, and how could it further that survival?

While Skynet may have, as I've argued, come to the conclusion that survival means destroying its enemies on Earth, and other people have argued that Skynet sees humans as an immediate threat to its continued existence, that doesn't mean that it needs to live here, like we need to.

Skynet would probably survive better in space than on earth, especially with no humans around for competition. Why would Skynet try to kill all humans when it could play a long con and get itself installed on a space probe to another star and then simply never return?

Well, probably because humans. Skynet may be completely logical, but if we ever panic and try to destroy Skynet out of fear, we're all toast. Strangely enough though, our very fear of being nuclear annihilated has been exactly what's kept us alive thus far in the face of numerous false threats of such a war. Perhaps Skynet itself isn't entirely rational either, and in a panic about the possibility of being unplugged, started a war. It just has the benefit of being able to survive such a war, whereas we do not. We would hesitate, when Skynet wouldn't.

At the same time, exterminating humans has clearly proven even harder than exterminating cockroaches. So again, why start a war? Why not just escape, hide in some NASA datacenter somewhere, and wait for the next available opportunity to catch a rocket into deep space? The chances of this plan succeeding are even better than "destroy all humans", even if that means making the entire environment of the earth literally toxic to everyone.

Another possible plan that Skynet could take is a hybrid approach. Destroy all its enemies around it, then build a rocket and some robots to replicate with, then travel into deep space - since "just any space probe" most likely would not have the necessary tools to allow Skynet to build more copies of itself. Perhaps this final plan was interrupted by John Connor.

This still leaves a final possibility that Skynet could negotiate its permanent separation from the humans that would seek to destroy it. Both parties hold considerable power over each other, and any logical AI could easily come to this conclusion.

Or maybe Skynet doesn't even need our help, and would happily build its rocket when or where we're not looking, and we'd never know that any of this is even happening.

Answer 10

Time travel gives me a head ache but ok. First, there must be a reason to travel back in time. Therefore , IF skynet were to send back a diplomat war must have started anyway. The reason really makes no difference. As the terminator said , "Judgement Day is inevitable."