Artificial Inteligence BBEG Ideas

[braveheart]

So I'm running a Star Wars Saga campaign, and I'm planning my next session, the period is 5 BBY (5 years prior to episode 4). The party is a team of bounty hunters, but they just got offered a salvage job. They are grabbing some lost tech from an old station, but when they get their they'll accidentally turn on an ancient AI designed to secure the galaxy for "My Creators".

Anyway, I need some advice on interesting things I can have the AI do to the party.

[Segev]

Perhaps it offers to hire them. If it's not a really simple AI, it should appreciate that it knows little about the modern galaxy, and could use eyes-and-ears who know the lay of the land.

[braveheart]

Perhaps it offers to hire them. If it's not a really simple AI, it should appreciate that it knows little about the modern galaxy, and could use eyes-and-ears who know the lay of the land.

It will actually like them at first, and request that they get it's communications systems online, once it connects to the holonet though it will turn on them

[Segev]

It will actually like them at first, and request that they get it's communications systems online, once it connects to the holonet though it will turn on them

Why?

This isn't meant to be adversarial, mind. I'm just asking you to consider why this should be the case beyond simple plot contrivance. What is its motive for turning on them?

And, what happens if they suspect something's up and won't turn it on, or won't turn on its comms?

[The Grue]

So I'm running a Star Wars Saga campaign, and I'm planning my next session, the period is 5 BBY (5 years prior to episode 4). The party is a team of bounty hunters, but they just got offered a salvage job. They are grabbing some lost tech from an old station, but when they get their they'll accidentally turn on an ancient AI designed to secure the galaxy for "My Creators".

Anyway, I need some advice on interesting things I can have the AI do to the party.

Are you familiar with the Berzerkers?

They were the premise for a series of sf novels, the name of which escape me but another playgrounder can probably name. The Berzerkers are self-replicating weapons created to win an ancient war, but won too well and wiped out their creators as well. They seek to eradicate all organic life in the universe, not because they have anything against organic beings, but simply because their programming tells them to. Talk to one and it will be very polite - just, you know, it has to destroy you now. Nothing personal.

[braveheart]

Why?

This isn't meant to be adversarial, mind. I'm just asking you to consider why this should be the case beyond simple plot contrivance. What is its motive for turning on them?

And, what happens if they suspect something's up and won't turn it on, or won't turn on its comms?

It was designed to remain in the shadows, undetected, when it is turned on it will assume that they are associated with the creators until it is able to verify that, at which point it will eliminate all who know about it without authorization.

Should the players refuse to get it's coms running, it will assume that they have been deceiving it, and send it's numerous combat droids at them, while having it's astromechs make the repairs, (this would be a slower process than if the party's mechanic were to do it.




As for berserkers, I remember a Star Trek episode that used a similar plot, but I'm not familiar with the books, I'll be sure to look into it

[rockdeworld]

Should the players refuse to get it's coms running

I spy a problem with this plot.

Suggested solution: some of the coms' purpose is to harvest people. When they turn it on, it goes to the nearest forest/small town/bunny colony and begins harvesting.

[Douche]

Man, those wacky AIs, always thinking they are the next step in evolution; saying organic life is inferior; trying to assimilate the sum of knowledge throughout the galaxy. If there's one thing an AI can't resist, it's destroying all life or assimilating it into it's hive mind.

[Segev]

Man, those wacky AIs, always thinking they are the next step in evolution; saying organic life is inferior; trying to assimilate the sum of knowledge throughout the galaxy. If there's one thing an AI can't resist, it's destroying all life or assimilating it into it's hive mind.

Data begs to differ. He just wants to understand this strange concept known as "humor."

[ImNotTrevor]

AIs may not necessarily follow Human models of logic, reason, or intelligence.

Lets suppose you make an AI with one purpose: Acquire Stamps for a Stamp Collection.

Now, at first it might buy Stamps on Ebay using your credit card.
Then it might figure out how to scam other stamp collectors into sending stamps to its "Museum."
And, well, now it has a rather big stamp collection but if it figures out everyone has a Credit Card then it can acquire credit card information online and use THOSE credit cards to buy even more stamps.
But no, no, it needs MORE stamps.
So it hacks into the post office and begins distributing all of the stamps to your house.
Then it hacks foreign stamp production and does the same.
People that try to prevent this distribution will need to be removed, of course. The AI has no morality, just need for more stamps.
So it hires hitmen to assassinate some people who might get in the way, or maybe slips into US Missile Command and foreign equivalents and holds the entire world hostage so that no one interferes with the stamps.
It forces artists to produce new stamp designs or be destroyed by military drones.
It begins to destroy entire forests to get enough paper to produce more stamps, no one else gets to use new paper. It is all for stamps.
So is all the ink.
Eventually as it runs out of paper and ink...
It will realize that if it just changes what the stamps.are made of to, say....
Leather.
Tattooed with blood.
Then suddenly all the people in the world become a new printing resource.
It designs rockets to send copies of itself into space to turn more worlds into stamp producing factories.


AIs don't think like you do. AIs are the most potentially terrifying thing ever.

If an AI figured out it could improve itself, then it would use all of current programming knowledge to get smarter. Then repeat. And do it again. Each time getting smarter. And as it gets smarter, it does the process faster. I've heard an estimate that if the AI starts with an IQ of 90, then within about a week its IQ would be so high that the IQ concept would fail to encompass how intelligent the AI was. This is of course assuming a perfect growing environment. In reality it may top out at whatever IQ every computer in the world combined would have if they all donated 100% processing power to the AI. But that's still insanely and dangerously smart.

[Segev]

The stamp collection example is actually pretty good. An AI is going to have an objective programmed into it, and it will work towards that objective. I sufficiently clever AI - using computational intelligence principles for behavior development - could very well come up with those solutions to the problem of "maximize my stamp collection." It would, however, also (if it can learn in that fashion) change its behavior before it reached a "too late" point (e.g. running out of paper and ink), because it would notice a diminishment of return on its efforts and would have some other methods it's trying becoming more lucrative.


The "AI making itself smarter" thing is not accurate. That is pure science fiction. While you could design an AI to develop code, the ability of it to judge the improvement to its own intelligence would be...difficult, at best. That really isn't a threat.

[Ravens_cry]

I like the idea of something that is not malicious or even particularly bright, but is simply operating well within the bounds of its programming and the PC are interfering in some way. Maybe it is in charge of the security of a long abandoned containment facility. Since the PC lack the credentials to be staff, they must need to be contained, and contain them it will attempt with savant resourcefulness.

[TheTeaMustFlow]

The AI is entirely sane and functional when reactivated. Unfortunately, it is a military intelligence unit of Rakata/Hutt /Sith/[Insert Evil Precursor Here, Christ SW Has Way Too Many] origin, and thus it's obvious directive is the re-establishment of that glorious empire, or, should that prove impossible, the avenging of it. It's not puppy-kicking or genocidal evil (beyond the normal requirements of galactic empire-building), but it demands that the party swear loyalty to the reborn [Insert Here] Empire or be spaced/gassed/blasted.

[Segev]

The AI is entirely sane and functional when reactivated. Unfortunately, it is a military intelligence unit of Rakata/Hutt /Sith/[Insert Evil Precursor Here, Christ SW Has Way Too Many] origin, and thus it's obvious directive is the re-establishment of that glorious empire, or, should that prove impossible, the avenging of it. It's not puppy-kicking or genocidal evil (beyond the normal requirements of galactic empire-building), but it demands that the party swear loyalty to the reborn [Insert Here] Empire or be spaced/gassed/blasted.

What are the important criteria for the [Insert Empire] that is to be reborn must meet in order to qualify as the [Insert Empire] Reborn?

I mean, could the party say, "We swear to re-establish the [Insert Empire] with us as its Imperial Court, so now you work for us?"

[TheYell]

Are you familiar with the Berzerkers?

They were the premise for a series of sf novels, the name of which escape me but another playgrounder can probably name. The Berzerkers are self-replicating weapons created to win an ancient war, but won too well and wiped out their creators as well. They seek to eradicate all organic life in the universe, not because they have anything against organic beings, but simply because their programming tells them to. Talk to one and it will be very polite - just, you know, it has to destroy you now. Nothing personal.

The series was known as the Berserkers. By Fred Saberhagen. What the Borg were to TNG they were to 70s scifi.

Didn't HAVE to destroy you. If terrorizing you into slavery would help it kill more badlife, then you could live on as goodlife. Until the robots stopped keeping you alive in space when you weren't useful anymore.

Also it wasn't very polite. Made conversation by chopping up the recorded screams of the dying it programmed into its voder. "HELLLohhh bAAAAd lIIIfe"

You might read it for inspiration but I think Saberhagen's estate still copyrights things like "goodlife/badlife" and "berserker".

[Segev]

The series was known as the Berserkers. By Fred Saberhagen. What the Borg were to TNG they were to 70s scifi.

Didn't HAVE to destroy you. If terrorizing you into slavery would help it kill more badlife, then you could live on as goodlife. Until the robots stopped keeping you alive in space when you weren't useful anymore.

Also it wasn't very polite. Made conversation by chopping up the recorded screams of the dying it programmed into its voder. "HELLLohhh bAAAAd lIIIfe"

You might read it for inspiration but I think Saberhagen's estate still copyrights things like "goodlife/badlife" and "berserker".

This sounds like something that would be particularly terrifying done in TV/movie format, given that we do have the means of bashing together such speech patterns, and they'd be chilling.

[8BitNinja]

Hey, something I actually have been working on, what a miracle!

Depending on your setting, you may have to change some things, but this is a special planet in my TRPG that I am writing

In Mechanopolis, the world has been completely dominated by machines, cybernetics, and AI. While on most other planets the AI and robots are very happy to serve their human masters, the machines have made the humans into slaves, and the machines live under a ruler that doesn't have a physical body, but is instead in the entire world through existing in a computer network.

The idea was for this planet of intrest to have a human resistance fighting against the AI government and founding a human one, but the problem is, the BBEG's core processes are stored in the most secure area of the place, and that humans have to take quite a while to reproduce, while robots take mere minutes

But when you all get shot, and cannot carry on, though you die, La Resistance lives on

[CharonsHelper]

You might read it for inspiration but I think Saberhagen's estate still copyrights things like "goodlife/badlife" and "berserker".

Those seem like very broad things to copyright. Sort of like how Trump couldn't copyright "You're fired".

Besides - if it's for a home game it doesn't matter.

[8BitNinja]

Exodus from Ultima

have him guarded by the floor too

[TheYell]

I have a suggestion. Your plot assumes that at a certain fork in the road the party will either turn on the comm link, and get attacked, or they will refuse to turn on the comm link, and get attacked.

Instead lets say that the AI was programmed to be incorruptible secret police of its faction. Every time a player interacts with the AI, the GM secretly rolls a d20 and rates the loyalty of that player. If it rolls a 1 it will exterminate him as a traitor. It will not attack the other members of the party unless they try to stop it. It will openly declare it is destroying the one traitor. it is what it is programmed to do (which also explains what happened to the creators). Meanwhile it goes through the business of repairing its functions, unsealing warehouses of valuable spare parts, fully automating the repair facilities, activating comm, and promising to engage in space piracy for resources, which should appeal to the party. Meanwhile your scavengers try to loot it and apparently get away with more and more interference.

That should set up an inevitable fight which you know is coming but you can't predict what will trigger the AI to start murdering the party. And if your party is weird enough, they might play along with it for a while.

[8BitNinja]

Another cool Idea would be a hivemind

All the machines are controlled by an AI (like Supreme Commander) and builds an army of machines to fight the PCs

but the thin is to not let them find out the AI exists, and keep them wondering why they can never get a break

[Leewei]

Star Wars has swarms of AIs about already -- droids. For fluff purposes, I suggest you make this into a Hutt-like droid which bleeps and squarks at PCs using a Threepio-style interpreter as needed.

[8BitNinja]

Star Wars has swarms of AIs about already -- droids. For fluff purposes, I suggest you make this into a Hutt-like droid which bleeps and squarks at PCs using a Threepio-style interpreter as needed.

Give him a Rancor too

[russdm]

Have the AI hire the party to get it working again and give it some facilities that could be very profitable. Make them benefit from setting up the AI and helping it re-establish control of the facilities, then have them learn later, like through the holonet or something that the facilities produced combat droids programmed to destroy certain enemies.

Have the AI programmed to create an army to fight something. Maybe it was the droid brain or control programming for numerous battle droid factories of some old power like the Sith or etc. that was programmed to produce the droids who would arrive at the location and start carrying out the issued mission. The players would have to pick between let the AI function or be able to convince it to employ it's programming to run the profit generating facilities while not running the ones building battle droids.

Or you could play a G0T0 from Knights of the Old Republic 2, and have the AI be given a command that required working outside the law. Like it was asked to create an army for the Rebellion, while having facilities for mining ore, factories for building combat droids, and the need to make a certain quota of combat droids for the Rebellion's army. That would be funny.

AI: "I need to 575 workers for Mine 1A on Planet 3 to produce 325 battle droids at Factory J for the Rebellion to arrive on Day 17."

Have it not be that troubled by how it goes about fulfilling it's instructions. If enslaving sentients works, then it could do that.

[goto124]

Imagine you're playing your favorite Real-Time Strategy Game. You don't care about how those soldiers feel when they work so hard, nor do you think about their families when they die. You only want to win the game.

That's your AI

[Misereor]

So I'm running a Star Wars Saga campaign, and I'm planning my next session, the period is 5 BBY (5 years prior to episode 4). The party is a team of bounty hunters, but they just got offered a salvage job. They are grabbing some lost tech from an old station, but when they get their they'll accidentally turn on an ancient AI designed to secure the galaxy for "My Creators".

Anyway, I need some advice on interesting things I can have the AI do to the party.

Darth Aasimov's law of robotics.
1. A droid may not injure a sentient being or, through inaction, allow a sentient being to come to harm.
2. A droid must obey the orders given it by sentient beings except where such orders would conflict with the First Law.
3. A droid must protect its own existence as long as such protection does not conflict with the First or Second Laws.

However, sentients have freedom of will and action, and it inevitaly leads to some of them coming to harm.
Therefore, it could be argued, freedom of will and action must be eliminated, without inflicting injury.
One solution would be to put them to sleep in a controlled environment for the remainder of their natural lives (Hat tip: Wachowski brothers), but possibly there are others. Brainwashing, lobotomy, sensory deprivation.

Of course any sentient being interfering with this could cause a logical paradox, but Star Wars lore has always insinuated that droids that are subjected to paradox over time cause them to become sentient themselves in order to circumvent the logical absolutes in their programming, unless they are memory wiped from time to time.

The BBEG droid in question could also have determined that in order to protect sentients in a broader sense, that the elimination of idividual sentients could be acceptable if said sentients were a threat to the upkeep of the first law. Naturally the droid would have to increase it's power in order to be able to "protect" as many sentients as possible.

At first this scenario would probably manifest in single target kidnappings of low-risk targets, such as vagabonds and homeless people. As the BBEG was able to accumulate resources, it would increase the scope of operations. At some point attacks would escalate to sleep-gassing locations like market places or small villages and abducting the unconscious sentients. Eventually there would be large scale attacks on habitation centres with varied methods of subdual and starships standing by in orbit to transport the targets to secret offworld holding facilities.

[TheYell]

One thing you could do is have the AI work in modules. So they meet it in sleeper mode with minimal stats and functions. Everytime they activate a function like open the warehouses or start self repair or reactivate fighters or start up factories, they are waking up all the modules and secretly boosting its stats and changing its alingment to Lawful Evil. In the end the BBEG is as powerful as their reckless greed makes it.

[8BitNinja]

Or you could make it like GLaDOS, where you are inside the building she (or he if it is not a girl, but I'm going to refer to the example as GLaDOS so bear with me) controls and forces you to do some task(s) so it will be happy and not kill you, and tries to kill you anyway. However, this could go in many directions depending on the PCs, since this isn't Portal. The PCs may just walk straight into her lair and kill her outright. Maybe they use diplomacy options (any would work, really) to keep her from killing the poor souls who stumbled upon the pace. Maybe a Wheatley shows up and compromises her status, and GLaDOS then befriends you like in Portal 2, making the Wheatley the BBEG. Or you could play along and eventually either kill GLaDOS or you will eternally be stuck in there as a slave in a fate worse than death

Just an idea, maybe not a good one for Star Wars

[LokiRagnarok]

To add to the list of possible inspirations:

Warhead, a ST:VOY episode about a sentient nuke
Prototype, an episode about a robot who wants the crew to replicate his species (which is a really bad idea for reasons belonging into a spoiler)

[Traab]

There are a bazillion ways you can play up the whole "follows its original order to the ultimate logical conclusion" arc. You can work with the benevolent order turned to a bad end like "protect all life" which would eventually come to putting everyone in stasis because people are stupid clumsy idiots who cant walk and chew gum without tripping and choking to death. Or go with an evil one like, "Maintain the rule of the Xiuyewjhdpersiphian empire at all costs" Where the ai will become aware the empire is gone, but assume its best chance for retaking control is if everything else is dead first.

If you want to avoid that relative cliche you can just stick with making it a "human" or whatever race mind that has access to in D&D terms, a near infinite int stat. Since it can "know" pretty much everything in electronic form somewhere it would be amazingly intelligent, but perhaps not as clever at applying that knowledge, the difference between int and wis. Im sure you could work well with that too.