Ship AI communication

[c_osborne]

I have a question that is bugging me, and won't have an official answer, but I'm hoping for some ideas.

The PCs discover a high-tech ship, which runs a Model-9 computer with high-AI. Once they power it up, the AI wakes up.

Issue #1: How would the ship react to "aliens" (PCs) on board? Assuming it is smart enough to realize it may need their help to survive, I'm thinking it might 'cooperate' at least a bit.

Issue #2: How would the ship communicate with the "aliens"? Would you treat it as just another NPC who doesn't speak the same language?

Any ideas would be greatly appreciated.

 

[Lenadh]

It would really depend on who created/programmed the AI. Results could be anything from friendships to **EXTERMINATE! EXTERMINATE!**

Yaskoydray had a ship that had a computer that was also psionic on some level (at least for interfacing), so one option would be for a person that tries to use the ship to get thoroughly mind-probed. The system then learns his language, and maybe some good (or bad) things about the current state of the universe...

 

[Icosahedron]

If your TU has aliens, presumably translator chips/programs will be plentiful, so it shouldn't take long for the AI to learn your language. It would probably pick it up fairly quickly just by listening to crew conversation (depending how far removed the two languages are).

As a free thinking entity, the AI would be able to make up its own mind how to react, depending on the circumstances, but as already suggested, its thought patterns would be rooted in the culture of its creators. However, perhaps a lot would depend on how long it had been sentient. Many humans reject their birth culture and the ideologies of their parents as they grow and learn. I would suspect that the first reaction of a highly intelligent entity finding itself in a new situation would be to learn. I think *INPUT* is more likely than *EXTERMINATE* unless the machine is psychotic. More than likely, though, its next imperative would be *ET PHONE HOME*.
It would certainly have its own agenda, and its reaction to humans would probably depend on what it discovered about human culture and behaviour from the examples available to it.

 

[mike wightman]

If the ship is damaged or out of fuel then the AI will realise the PCs may be necessary to effect repairs or refuel.
So logically it is likely to try to learn to communicate - listening in on the PC communications etc.
Its attepts at return communication could be interesting...

 

[Plankowner]

Imagine a ship trying to communicate to the players using flashing lights in Morse Code...

Would the PLAYERS even recognize what was going on? Could the characters?

 

[c_osborne]

That's where the trouble lies. "Home" doesn't exist anymore, and the ship will be in bad need of maintenance/repair. But the ship won't be able to easily communicate with the PCs either, so I'm just struggling with the likely sequence of events. A lot will depend on how the PCs decide to act.

 

[drmike]

I've been toying around with an AI developing and learning as the characters adventure. To represent this the AI has "aged" as it has learned. Right now it is an adolescent. It's been fun to see the players reactions when the AI has a temper tantrum and refuses to communicate or just plain does what it wants for a couple turns.

 

[Sifu Blackirish]

Now THAT is an interesting concept...
In my mind's eye, I see some scenarios...

1) the computer has taken severe damage, and has to learn all over again. It has essentially been 'reborn' when the PCs board.

2) if it had access to some core memories (source files before mission specific programming), that might be enough for it to recognize what it is, when it reboots.

3) it might mis-identify one of the PCs due to their service symbol being close to a friend/enemy of its builders. There could also (and/or) be an emotional attachment, with romantic complications...

4) it has been awake all this time and is now suffering from from some sort of psychosis. Shades of 'Dark Star' here.

You could set up a 'decision tree' for the computer actions, if it is only partially functioning. Throw a reaction table and a patron action table into a blender and voila! Instant decisions.

 

[kafka47]

We all know how HAL & SAL evolved as a character in Clarke's novels. However, an AI could have very confused motivations, if you look at the Who episode - The Girl in the Fireplace, as it searches to complete itself.

I could envision a scenario where the AI is benign but was attacked by a computer virus, in fact its own Anti-Highjack program that got corrupted somehow giving a very real Paranonia RPG feel to the ship.

It could recognize the characters as the rightful heirs of a long dead Pocket Empire and try to persuade them to rejoin the Crusade against the upstart Sylean Empire...leading to particular complications with current Imperial authorities...

#2 Could the computer maybe communicate psionically? Using holographic projectors would give the impression of ghosts.

 

[mike wightman]

IMTU the reason the Kinunir AI went crazy is a result of it also being a test-bed for the deyo transponder...

the next time I run the adventure I'm going to have the Kinunir computer learn from the mistake of killing the crew and actually work with the PCs, at least at first...

 

[whartung]

If it's truly an advanced AI, just treat the ship as a person locked in a box. How would a person locked in a cell try and communicate with the outside world, etc.

If you're using TNE, draw two cards for the AI motivation, etc.

Basically just treat it as another NPC, an NPC that happens to be locked in to the bridge of a ship.

 

[Keklas Rekobah]

Issue #1: How would the ship react to "aliens" (PCs) on board? Assuming it is smart enough to realize it may need their help to survive, I'm thinking it might 'cooperate' at least a bit.

Issue #2: How would the ship communicate with the "aliens"? Would you treat it as just another NPC who doesn't speak the same language?

-------

Assuming that the AI and the 'aliens' speak the same languages, there are some things to consider:

1) An AI, while possessing a sophisticated stimulus-response decision tree, can not reason beyond its database. The exception would be those neural-net types that can actually learn by induction. Thus, if the AI possessed the equivalent of Steward-8 and nothing else, it would not necessarily be capable of operating a starship, even if its equivalent IQ were 300+ and it had access to the entire ship's library, unless it could both learn and understand the data.

2) An AI may only be able to effectively communicate within the context of topics related to its intended functions. Consider a nerd; he or she may go on for hours about 'Magic: The Gathering' or the musical styles of Enya, but beyond this, their conversational abilities will be full of references to their special interests. ("When will lunch be ready? Who knows? Only time...")

3) Read this article on Asperger's Syndrome, and apply the concept of a person with 'Highly Functional Autism' to your AI PC/NPC. The key phrases are "normal intelligence and language development" and "marked deficiencies in social and communication skills."

4) Communication should not otherwise be a problem, once you figure out what the mode of communication is ... "Okay, try this: tug your left ear twice, blink, and hop on your right foot ... good ... now that we have life-support running, let's try to get the m-drive on line... take your right hand and grab your... GESUNDHEIT! Well, now we know the missle launchers work..."

5) Also, keep this in mind, especially for AIs running military vessels and high-TEC dictatorship-run worlds: "Worlds governed by artificial intelligence often learned a hard lesson: Logic Doesn't Care." -- YIN-MAN WEI, "This Present Darkness: A History of the Interregnum", CY 11956 (from the TV series "Andromeda")