|
Purely a social question here if you were the only crew of a starship (not necessarily you, you, pick a character if you like) and you invested in a Artificial Intelligence would you rather have a VI or a true AI?
Personally I find it very character dependant which I prefer at any time. Probably because if I had to choose one I'd get in real life I find I keep flip flopping on which I prefer. An AI is a genuine person they'll offer suggestions, have desires and provides a level of interaction a VI can't. On the other hand part of that interaction is they may get bored or worse start disliking you and contemplating murdering you in your sleep (Imagine an AI with Arnold Rimmer's personality as your only companion as he is at the start of Red Dwarf). A VI has no feelings they'll simulate friendship but at the end of the day they're not going to change, not decide they want to strand you on a planet while taking the ship and you can always flip the personality depending on your mood today I want a sean connery style VI tommorow a John Cleese and on Wednesday I"ll base it on Goldie Hawn from house sitter. Meanwhile if they're advanced enough you can probably fool yourself into thinking they have genuine feelings of friendship towards you.
| 1 person marked this as a favorite. |
One of my players wants to romance the ship.
In that case, an AI is pretty much required for it not to be creepy.
A good example of VI would be Joi from Blade Runner 2. The Doctor from Voyager is a better example of a VI that became an AI. There's also Gene Roddenberry's Andromeda which got a good example of an AI and the AI's android avatar.
Actual me, personally, I would always go with VI (assuming you mean virtual intelligence) for any tool I'm using or anything I want to rely on.
I don't want my Starship suddenly thinking my choices are wrong and it decides to open all the door on the ship (including the air locks) and start trying to purge itself of the organic infestation aboard it.
Nah. I don't trust actual real AI. Especially not with so much control over things I use/need. We all work with people, imagine that one coworker you don't like very much (or maybe you have one you know doesn't like you for some petty reason) and now imagine they could turn off oxygen just for you. While you're asleep. And you just never wake up.
Actual AI is super dangerous. Because it's all the fickleness of a person, with potentially the ability to launch a nuclear strike against you.
I could see doing a EDI from Mass Effect situation where you have an AI on board the ship that doesn't have control or access to any important system of the ship, but is available in a sort of advisory role and for it's advanced calculation abilities.
One of my players wants to romance the ship.
In that case, an AI is pretty much required for it not to be creepy.
I got to be honest, to me it's creepy with or without the AI.
Of course, if you're on a ship and that one co-worker decides to override the environmental controls and turn your oxygen off, it's pretty much the same situation.
The AI is just another co-worker and in a critical situation like space or like adventurers get into any of them can kill you pretty easily.
I assume that at least on most ships life critical systems are at least locked to specific personnel, like the captain or head engineer/mechanic.
Joe the red shirt from the landing team isn't going to have permissions to turn off the oxygen because you got his friend John killed last week on an away mission.
Sure, there are ways for him to hack it. And sure you could attempt to restrict the AIs security access to things....
But I prefer to not even introduce the option by not having an actual AI running the ship. Hell for the most part you don't need an AI to run the ship. What can an AI do that an equal advanced VI couldn't?
Basically just generate new ideas and having interesting conversation.
I don't need my ship to do that, that's why I have a crew.
A good example of VI would be Joi from Blade Runner 2. The Doctor from Voyager is a better example of a VI that became an AI. There's also Gene Roddenberry's Andromeda which got a good example of an AI and the AI's android avatar.
Andromeda is wierder then most examples. She has 3 seperate personalities. The android, the hologram, and the viewscreen image. They had the same memories and experiences, but were still seperate personalities of the same AI. I don't recall any of the other AIs have the same split, although some of the android avatars did if they didn't reconnect with the core AI for some time.
|
|
A good example of VI would be Joi from Blade Runner 2. The Doctor from Voyager is a better example of a VI that became an AI. There's also Gene Roddenberry's Andromeda which got a good example of an AI and the AI's android avatar.Andromeda is wierder then most examples. She has 3 seperate personalities. The android, the hologram, and the viewscreen image. They had the same memories and experiences, but were still seperate personalities of the same AI. I don't recall any of the other AIs have the same split, although some of the android avatars did if they didn't reconnect with the core AI for some time.
Interesting we do have quite a range of "AI" in fiction from Andromeda and the 3 personality AI, Holly in Red Dwarf who's gone senile, Cassandra also from red dwarf who's generally friendly but makes people uncomfortbale because she know's how they'll die to Data from Star Trek with and without emotions and other examples some of whom don't even realize they're an AI.
This is part of why I tend to flip flop if you have a humanish personality and mind set (Rimmer from red Dwarf, Doctor Schweitzer from Star Trek, Rachael from blade runner it'd be no different than having another crew member. I'd trust that AI to be chief engineer, CMO, captain or the like as much as any normal human. They are the same in all the important ways. Then you move onto AI's like Sam from Mass Effect or Cortana from Halo. They're close but have some fundamental differences, they're viewpoint just isn't the same and there's that question of "They're on my side now but for how long." that'd linger in the back of my head. Moving onto the other side of scale you have AI's like Ash from Alien that are fundamentally alien and can go insane if you don't handle them right which would be uncomfortable to be around. Finally we verge into the fully different mindset and hostile nature like Skynet from Terminator.
What kind of AI do we have in starfinder? Is it closer to a real person (Digital nature notwithstanding) or is it an emotionless, alien mindset?
Sure GM discretion but still the nature of the AI affects how much you'd trust it.
EDIT
Thinking about it technically Dawn from Buffy the vampire slayer is an AI too.
What kind of AI do we have in starfinder? Is it closer to a real person (Digital nature notwithstanding) or is it an emotionless, alien mindset?
Considering that androids, anacites, SROs etc. exist I don't see why ship AIs would be any less evolved.
The question is, do you want your ship to have an AI? What is the benefit? And I already assume that ship AIs are already made in a way that they are ok being a ship with very limited social circles, ways of entertainment or other ways to interact with society (or you have ship AI poker sessions like the hotels in Altered Carbon)
Consider that VI run freighters that spontaneously become full AIs tend to join pilot unions and demand better wages, or split with their former owners entirely, well I'd say it varies AI to AI.
SF androids at least canonically have souls. I'm not if that's true of AIs.Considering that androids, anacites, SROs etc. exist I don't see why ship AIs would be any less evolved.
What kind of AI do we have in starfinder? Is it closer to a real person (Digital nature notwithstanding) or is it an emotionless, alien mindset?
"Rarely a starship VI's sentience develops to the point where it attracts or generates a soul and becomes a true artificial intelligence (or AI)"
Canonically AIs have souls if they are full AIs. SROs have souls just like androids, and whether anacites have souls vary from anacite to anacite.
Of course, if you're on a ship and that one co-worker decides to override the environmental controls and turn your oxygen off, it's pretty much the same situation.
The AI is just another co-worker and in a critical situation like space or like adventurers get into any of them can kill you pretty easily.
I assume that at least on most ships life critical systems are at least locked to specific personnel, like the captain or head engineer/mechanic.
Joe the red shirt from the landing team isn't going to have permissions to turn off the oxygen because you got his friend John killed last week on an away mission.
Sure, there are ways for him to hack it. And sure you could attempt to restrict the AIs security access to things....
But I prefer to not even introduce the option by not having an actual AI running the ship. Hell for the most part you don't need an AI to run the ship. What can an AI do that an equal advanced VI couldn't?
Basically just generate new ideas and having interesting conversation.
I don't need my ship to do that, that's why I have a crew.
So the AI is a person, just like the rest of the crew. The AI has access to critical life support just like some other members of the crew. If you can trust them, why not the AI?
The AI is a person, but the Ai of a ship had drastically more power over things related to the ship, than any one person does.
Let's not imagine the small ships that PC players are likely to have, and instead think about the very large ship that are available that are likely to have crews in the 100+ range. Imagine if everyone of them could open all the air locks and turn off the air. I'd be terrified. I wouldn't want to be crewed aboard such a ship. On a large ship I'm imagining the people who are allowed access to critical systems is restricted to necessary individuals who have proven themselves over time. It doesn't mean they're not susceptible to doing bad things, but at least the probably haven't before.
So why trust other people on the Starship? Well I can't do it all by myself, and their restricted access means their threat to catastrophically hurt me is manageable. But restricted access doesn't really work on an AI that is your ship. It's workable with other AI's who are in a body, which will presumably have restricted access.
But there's simply no advantage, from my perspective, by having a ship with a true AI. All I need is a good VI.
I got security crews to respond to crazy murderous people and a VI that can monitor things, and alert the right people if something moves out of specific ranges. VI is really good at waiting and watching.
And no, the AI isn't proven until you've run missions with it. But I'd rather not give it a chance in the first place because it's not worth the risk in my view. I've got nothing to gain and lots to lose by choosing an AI over a VI. If the AI wants to join the crew, I'll let it on in an SRO body without access to anything important until it proves it can be trusted.
Honestly, if I could I wouldn't have anyone else adventuring with my but it's impractical and likely to get you killed. So you just need to keep it the minimum of people that have proven they can be trusted not to suffocate you by filling your cabin with CO2 in your sleep.
The part that you really don't seem to get is that there is no benefit to me, to let an AI run the ship. There's is very little an AI could do that an equally advanced VI couldn't.
|
|
What kind of AI do we have in starfinder? Is it closer to a real person (Digital nature notwithstanding) or is it an emotionless, alien mindset?
Considering that androids, anacites, SROs etc. exist I don't see why ship AIs would be any less evolved.
The question is, do you want your ship to have an AI? What is the benefit? And I already assume that ship AIs are already made in a way that they are ok being a ship with very limited social circles, ways of entertainment or other ways to interact with society (or you have ship AI poker sessions like the hotels in Altered Carbon)
Mainly for me its a matter of perception. I see the AI if properly developed as a person. They have hopes, dreams, plans. They might be working as a ships crew but have their eye on earning their way into that prestigious city management job. They wont have ethics circuits any more than organic beings do they'll have their own ideas of right and wrong. A VI on the other hand only acts and the difference between a VI who looks after the crew vs one that tries to upgrade them into a better being is that circuit board over there. I know its not entirely reflected in the rules which is why I'm not sure if the AI's ingame are that advanced. However to me a true AI would be like an android player. They'd have a personality, class levels and so on. They'd have an advanced brain as the core of their system so you can't just pull a "ethics circuit" or a "skill expander module" out and change them. Altering an AI would be as difficult as altering a person or as easy. Unlike others here I actually find an AI if how I picture them more reliable and secure a being to have on board than a VI.
Like I said if I knew (or was running the game) 9 times out of 10 I'd prefer an AI, if I'm not sure how advanced they are I tend to be less sure. The other issue is because I'm difficult to deal with and I imagine the AI would be no more likely to want to hang around and deal with me than anyone else. On the other hand while a VI would I'm not sure I'd be comfortable trusting myself to a being who looks after me because code say's it should.
The main reason I'm not sure how evolved the AI's in Starfinder for players are is the rules. We have the mechanics "True AI" drone that doesn't seem to have any personailty just decision making and VI rules for starships that may become AI's but unless houseruled are just a bunch of skills. I'm not certain if the AI's players can get are as advanced as the AI's players can be. If you see what I mean?
As for them being comfortable being a ship (Ship who sang or city who fought are good examples of what it might be like though those are people with a physical disability so severe they're implanted into circuits as baby's so they can use their talents without dying, literally you have to have a high chance of dying as a baby to be a viable candidate). The flip side is the holographic emiters you can get that allows a simulation (with elemental stats) inside the ship and up to 30' outside it. So the AI could well be like Doctor Schweitzer on Startrek or Rimmer on Red Dwarf. Aware of the ship but able to take physical human, vesk, yosokai, cat form and interact with things on the ship. Eating, chatting, having their own room with things they buy, etc. I just wish you could get a remote emiter so they could move beyond that range and have true shore leave just visiting the places they want to go.
I got security crews to respond to crazy murderous people and a VI that can monitor things, and alert the right people if something moves out of specific ranges. VI is really good at waiting and watching.
And no, the AI isn't proven until you've run missions with it. But I'd rather not give it a chance in the first place because it's not worth the risk in my view. I've got nothing to gain and lots to lose by choosing an AI over a VI. If the AI wants to join the crew, I'll let it on in an SRO body without access to anything important until it proves it can be trusted.
Honestly, if I could I wouldn't have anyone else adventuring with my but it's impractical and likely to get you killed. So you just need to keep it the minimum of people that have proven they can be trusted not to suffocate you by filling your cabin with CO2 in your sleep.
The part that you really don't seem to get is that there is no benefit to me, to let an AI run the ship. There's is very little an AI could do that an equally advanced VI couldn't.
So you'd prefer something like EDI from Andromeda if you were getting an AI one that had built in restrictions on what she can access until you release them.
|
|
In addition to the above if I were going with AI I'd love something like Rimmer's hardlight hologram emiter that essentially allows him to be a fully interactive being eating, holding things, the whole 9 yards. Not just Red Dwarf either in Voyager they replicated working holographic lungs and holosuites are apparently real enough in feeling they may as well be flesh and blood for interaction. Add in magic and the only difference between an AI hologram and an organic/silicon being would be that if you kill the hologram their mind is still safe somewhere else (though I'm sure some unethical company would love to exploit that) and they get the worse of being affected by humanoid/construct things like an android does. Part of why I think a true AI should essentially be a character with class levels and personality rather than skill expander modules.
So the HMS Starfarer AI mind can use a remote emiter or the shipboard units to generate a holographic body that essentially functions as a lvl 15 (VI level 15 and lvl 3 holographic emiter BP cost rules) Envoy complete with skill ranks, skills and HP that will disrupt the hologram and destroy the emiter. When active though she can travel anywhere (within coms reach), eat, fight, drag a crew member to see a sporting event the team she supports is playing in, grumble about the lines for the girls restroom, not like scary movies much less being sent on the away team to the apparently haunted derelict ship, etc.
Even if you don't have that this is how I see AI's they are people good, bad and neutral. Which is why I trust them more than a VI. It's just that I can easily see them getting annoyed with me and turning in their 2 weeks notice before looking for another job/ship/computer system.
So you'd prefer something like EDI from Andromeda if you were getting an AI one that had built in restrictions on what she can access until you release them.
I got security crews to respond to crazy murderous people and a VI that can monitor things, and alert the right people if something moves out of specific ranges. VI is really good at waiting and watching.And no, the AI isn't proven until you've run missions with it. But I'd rather not give it a chance in the first place because it's not worth the risk in my view. I've got nothing to gain and lots to lose by choosing an AI over a VI. If the AI wants to join the crew, I'll let it on in an SRO body without access to anything important until it proves it can be trusted.
Honestly, if I could I wouldn't have anyone else adventuring with my but it's impractical and likely to get you killed. So you just need to keep it the minimum of people that have proven they can be trusted not to suffocate you by filling your cabin with CO2 in your sleep.
The part that you really don't seem to get is that there is no benefit to me, to let an AI run the ship. There's is very little an AI could do that an equally advanced VI couldn't.
Yes, I would accept an EDI situation where they AI is not connected to anything (well EDI is connected to sensor IIRC but that's it) but can monitor and make recommendations to a situation. But doesn't have any control over the ship systems....until the end of the series when EDI has proven itself and they decide to integrate EDI with the ship because they need the "overlord mind" to coordinate for the massive starship battle about to take place.
Also Legion was cool, but he also didn't have any control over the ship. I'm fine with AI. Just not being in control of my ship until I've personally developed a rapport with them to trust that they're not going to go "Murder all flesh bags!" on me.
Even if you don't have that this is how I see AI's they are people good, bad and neutral. Which is why I trust them more than a VI. It's just that I can easily see them getting annoyed with me and turning in their 2 weeks notice before looking for another job/ship/computer system.
And this is even more the meat of the problem. What if in the middle of the mission the AI decides "Hey this is stupid dangerous and I like existing" and decides it doesn't want to transport you to the evil lair of doom which is surrounded by space monster and undead pirates. Can you force it? Would you force it?
Wouldn't it just be easier if you didn't have an AI spaceship?
If an SRO doesn't want to go on the mission they can just not go. Even in the middle of transport worst case they stay on the ship and pick up the pieces of you when you get back. The spaceship doesn't really have an option to not go if you're in a rush. And if that spaceship is also an AI...
|
|
So you'd prefer something like EDI from Andromeda if you were getting an AI one that had built in restrictions on what she can access until you release them.
I got security crews to respond to crazy murderous people and a VI that can monitor things, and alert the right people if something moves out of specific ranges. VI is really good at waiting and watching.And no, the AI isn't proven until you've run missions with it. But I'd rather not give it a chance in the first place because it's not worth the risk in my view. I've got nothing to gain and lots to lose by choosing an AI over a VI. If the AI wants to join the crew, I'll let it on in an SRO body without access to anything important until it proves it can be trusted.
Honestly, if I could I wouldn't have anyone else adventuring with my but it's impractical and likely to get you killed. So you just need to keep it the minimum of people that have proven they can be trusted not to suffocate you by filling your cabin with CO2 in your sleep.
The part that you really don't seem to get is that there is no benefit to me, to let an AI run the ship. There's is very little an AI could do that an equally advanced VI couldn't.
Yes, I would accept an EDI situation where they AI is not connected to anything (well EDI is connected to sensor IIRC but that's it) but can monitor and make recommendations to a situation. But doesn't have any control over the ship systems....until the end of the series when EDI has proven itself and they decide to integrate EDI with the ship because they need the "overlord mind" to coordinate for the massive starship battle about to take place.
Also Legion was cool, but he also didn't have any control over the ship. I'm fine with AI. Just not being in control of my ship until I've personally developed a rapport with them to trust that they're not going to go "Murder all flesh bags!" on me.
...
Even if you don't have that this is how I see AI's they are
It depends entirely on how you view the AI I guess but in starfinder I feel its even more reason to have an AI rather than a VI given your views or better yet neither.
An AI may not have that degree of mutual involvement and trust you want but generally I view them as not being fresh out of the simulator. If your part of a society that can make new ones that might be the case though even there you do know what kind of person they are. They'll learn and evolve like anyone but their starting personality and views are defined e.g. doctor Schweitzer from star trek, Rimmer from red dwarf or SHODAN from system shock (run). However since in Starfinder a core part of AI vs VI is AI have a soul I can't see them being mass produced without ethical and spiritual issues. Hmmm I'm sure I recall one setting where AI's where created as children and had to grow/develop like a normal being before being given access to more information but I can't place it.
So since your likely to be recruiting one who's done it before I figure you can check up on them and be sure they wont break at a bad time. That is you don't know Xerxe's personally but they would have a resume, experience, references and want to work in this environment. Just like your recruiting a new crew member. Yes they may break or things may change but its not any more likely than any other veteran having that issue.
A VI on the other hand going by starship operations could evolve to an AI at any time. You could be in the middle of say the explorations at the start of dead suns and just as your getting to the sunrise maiden to escape the VI on board becomes an AI and goes "Stuff this for a game of soldiers I'm out of here." and runs for it leaving you to die.
That said I think in your case as you say you'd be better off without either VI or AI. Given that you can't get an AI with restricted access until you've built that level of trust. For me as I said I do tend to flip flop on which I'd prefer at any given time. Of course I can see a difference between ship with mininimum 6 crew/maximum 20 having an AI with full access to everything generating trust issues and the standard PC one in starfinder where any single person can control all ship functions due to minimum 1 crew trusting one.
Hmmm we do know there are biomorphic controls so only given individuals can access the ship, firewalls and the like which added to VI's only being able to perform one role (pilot, gunner, science officer) at a time may mean you can lock particular access for particular crew AI or not.
A VI on the other hand going by starship operations could evolve to an AI at any time. You could be in the middle of say the explorations at the start of dead suns and just as your getting to the sunrise maiden to escape the VI on board becomes an AI and goes "Stuff this for a game of soldiers I'm out of here." and runs for it leaving you to die.
That said I think in your case as you say you'd be better off without either VI or AI. Given that you can't get an AI with restricted access until you've built that level of trust. For me as I said I do tend to flip flop on which I'd prefer at any given time. Of course I can see a difference between ship with mininimum 6 crew/maximum 20 having an AI with full access to everything generating trust issues and the standard PC one in starfinder where any single person can control all ship functions due to minimum 1 crew trusting one.
Hmmm we do know there are biomorphic controls so only given individuals can access the ship, firewalls and the like which added to VI's only being able to perform one role (pilot, gunner, science officer) at a time may mean you can lock particular access for particular crew AI or not.
I don't personally see VI as being likely of evolving. Lke it should be technically possible, but incredibly rare. VI is Google Home service just you know it's the future so it actually gets the commands correct and has access to a wider range of things it can do. Like "Hey Google, calculate the intercept course for that asteroid by not using our thrusters and instead using a gravity slingshot so we can surprise the pirates on it".
If VI did run a high risk or spontaneously becoming true AI then I wouldn't one on the ship either. But I'm assuming the VI is like the ships computer from Star Trek, which can do all sorts of awesome stuff with math, but isn't close to a True AI.
|
|
A VI on the other hand going by starship operations could evolve to an AI at any time. You could be in the middle of say the explorations at the start of dead suns and just as your getting to the sunrise maiden to escape the VI on board becomes an AI and goes "Stuff this for a game of soldiers I'm out of here." and runs for it leaving you to die.
That said I think in your case as you say you'd be better off without either VI or AI. Given that you can't get an AI with restricted access until you've built that level of trust. For me as I said I do tend to flip flop on which I'd prefer at any given time. Of course I can see a difference between ship with mininimum 6 crew/maximum 20 having an AI with full access to everything generating trust issues and the standard PC one in starfinder where any single person can control all ship functions due to minimum 1 crew trusting one.
Hmmm we do know there are biomorphic controls so only given individuals can access the ship, firewalls and the like which added to VI's only being able to perform one role (pilot, gunner, science officer) at a time may mean you can lock particular access for particular crew AI or not.
I don't personally see VI as being likely of evolving. Lke it should be technically possible, but incredibly rare. VI is Google Home service just you know it's the future so it actually gets the commands correct and has access to a wider range of things it can do. Like "Hey Google, calculate the intercept course for that asteroid by not using our thrusters and instead using a gravity slingshot so we can surprise the pirates on it".
If VI did run a high risk or spontaneously becoming true AI then I wouldn't one on the ship either. But I'm assuming the VI is like the ships computer from Star Trek, which can do all sorts of awesome stuff with math, but isn't close to a True AI.
I know what you mean but that is a REALLY bad choice to use as an example. "Computer create an opponent capable of being data." computer creates a sentient AI based on Moriarty, Emergence where the enterprise possibly becomes sentient and gives birth, Doctor on Voyager, that singer on DS9, Wesley's Nanties, the service robots on a mining drill that may or may not be intelligent, Voyager in the original series. Star Trek is full of AI's cropping up suddenly throughout the series.
EDIT
Thinking about it the TNG episode Emergence is probably a perfect example of your concern about AI. Ship becomes sentient, starts rebuilding itself and making decisions while locking out the crew from functions on the ship. Not malicious just wants to give birth and keep them all from dying horribly but still.
I know what you mean but that is a REALLY bad choice to use as an example. "Computer create an opponent capable of being data." computer creates a sentient AI based on Moriarty, Emergence where the enterprise possibly becomes sentient and gives birth, Doctor on Voyager, that singer on DS9, Wesley's Nanties, the service robots on a mining drill that may or may not be intelligent, Voyager in the original series. Star Trek is full of AI's cropping up suddenly throughout the series.
EDIT
Thinking about it the TNG episode Emergence is probably a perfect example of your concern about AI. Ship becomes sentient, starts rebuilding itself and making decisions while locking out the crew from functions on the ship. Not malicious just wants to give birth and keep them all from dying horribly but still.
That's fair. The average use of the computer on the ships in Star Trek is what I'm after. But there were specific episodes where dumb s&&$ happened that involved the ship's computer.
Although, I don't remember most of the examples you give I do remember the Moriarty/Data as Sherlock episodes and my real thought to that is: The computer shouldn't have been able to create an AI like that, IMO.
Star Trek has a bad habit (even within universe setting) of stretching what is reasonable or feasible.
The Doctor on Voyager was a separate specific program from the ship main computer. And is an example of a VI becoming an AI, though again I think it probably shouldn't be so easy.
I guess, if we extrapolate there is some position of advanced VI that is sufficiently close to being an AI that it can spontaneously convert. So ensuring that doesn't happen would be a matter of making sure you don't get to close.
And at least in the Lancer TTRPG which is an advanced SciFi setting they have the concept of recycling your VIs to make sure they don't become an AI. They basically put the existing VI into storage where a new copy of the original base programming can access information from it, but wouldn't theoretically have gained the "experience which might cause it to become an AI".
|
|
I know what you mean but that is a REALLY bad choice to use as an example. "Computer create an opponent capable of being data." computer creates a sentient AI based on Moriarty, Emergence where the enterprise possibly becomes sentient and gives birth, Doctor on Voyager, that singer on DS9, Wesley's Nanties, the service robots on a mining drill that may or may not be intelligent, Voyager in the original series. Star Trek is full of AI's cropping up suddenly throughout the series.
EDIT
Thinking about it the TNG episode Emergence is probably a perfect example of your concern about AI. Ship becomes sentient, starts rebuilding itself and making decisions while locking out the crew from functions on the ship. Not malicious just wants to give birth and keep them all from dying horribly but still.That's fair. The average use of the computer on the ships in Star Trek is what I'm after. But there were specific episodes where dumb s+@+ happened that involved the ship's computer.
Although, I don't remember most of the examples you give I do remember the Moriarty/Data as Sherlock episodes and my real thought to that is: The computer shouldn't have been able to create an AI like that, IMO.
Star Trek has a bad habit (even within universe setting) of stretching what is reasonable or feasible.
The Doctor on Voyager was a separate specific program from the ship main computer. And is an example of a VI becoming an AI, though again I think it probably shouldn't be so easy.
I guess, if we extrapolate there is some position of advanced VI that is sufficiently close to being an AI that it can spontaneously convert. So ensuring that doesn't happen would be a matter of making sure you don't get to close.
And at least in the Lancer TTRPG which is an advanced SciFi setting they have the concept of recycling your VIs to make sure they don't become an AI. They basically put the existing VI into storage where a new copy of the original base programming can access information from it, but...
AI's were a recuring story in star trek much like holodeck and transporter malfunctions going all the way back to the original series.
Still like I said I can see what your after you just want something like Mass Effect where there are deliberate coding blocks to ensure a VI doesn't become an AI while retaining star treks ability to interact. Which is probably fairly close to the basic starfinder computer. Me I want something like Red Dwarf's AI's Holly, Cassandra, Hudson, Rimmer. Good or bad at the end of the day they're people like anyone else.
I wouldn't be opposed necessarily an AI that I already trusted being placed into the computer of a Starship I would use, but I am opposed to being on a ship being run by an AI that I don't have a rapport with.
In the same way that I would have a problem getting on a ship with a flesh and blood captain I didn't know.
Although, for the sake of APs I would ignore that (if a character shared my concern) because it makes the adventures not work. Much like fn you made a character in PF1 that didn't want to travel.
|
|
I wouldn't be opposed necessarily an AI that I already trusted being placed into the computer of a Starship I would use, but I am opposed to being on a ship being run by an AI that I don't have a rapport with.
In the same way that I would have a problem getting on a ship with a flesh and blood captain I didn't know.
Although, for the sake of APs I would ignore that (if a character shared my concern) because it makes the adventures not work. Much like fn you made a character in PF1 that didn't want to travel.
Hmmm I think you might actually enjoy playing with my AI's then since I houserule them to be closer to the ones in star trek, red dwarf and the like.
Basically I want my AI's to be considerably different to a VI. So a VI is per the book (coding, modules, etc). An AI gets a soul and becomes a person complete with class levels, skill point allocations based on their interest, etc. The important thing for you is that since they're a person even if their in the ships computer as opposed to thier own minaturized portable one they can have their access limited by command codes like any other crew member. So A VI which is essentially a programmed interface to the computer has by necessity full access to everything with admin access even if it does restrict information to an individual. An AI which is a digital being interacts with the computer like a normal crew member.
You/AI computer VI-DATA/SYSTEMS.
or
You computer AI VI-Data/Systems
or
You computer AI Virtual Interface Data/Systems
the VI is tied to the data systems whereas the AI like you always interacts at one remove either holographically with a computer or with a virtual interface inside the computer.
So while say Chris the AI is on your ship computers that doesn't translate to full access. So he may be able to pull up sensors, power, weapons like any senior crew member but if he's say a lvl 12 Engineer hired to be the chief engineer (ships have holo emiters and I'm working on a mobile unit) then he'd have those command codes. That is he can give commands related to engineering he doesn't have say the self destruct codes or authority to power up the weapons. Likewise Clair the Yosokai hired to be the ships tactical officer has authorization over the ships weapons and internal defenses but she doesn't have the command codes to shut down the engines.
Even if Chris is not working as a holographic projection his access is still limited per his position and authority with other computer functions requiring different access to call up.
I guess the real question then is how the block in between the AI and the ship's system works.
As you mention, the VI is basically just an interface with some advanced assistant capabilities. It's capable of doing things like calculating a flight plan, because there are flight plan programs in the ship databanks and it's going to take your input and run them. The AI is actually doing it internal to itself.
I guess the part that I worry about it is, short of physical air gaps between the ship's computer and the AI (and even that may not be sufficient) how much do I trust that an AI is blocked from issuing commands to the ships computer.
I, as a human being who only has a beginner level understanding of computer programming, do not personally trust any amount of code to stop an AI which has decided to access the ship computer. If I recall correctly, which Cmd Shepard hooked EDI up to the rest of the ship it didn't have access to, it involved running cables throughout the ship and connecting it to EDI because it didn't physically have access to the parts it needed. That is the level of paranoia I have.
|
|
By a similar token, how do you keep the mechanic or technomancer you've just brought on board as crew from hacking the ship's computer?
This being the reason I trust an AI same as I trust anyone else skilled in computers. Either you extend that trust "this is a loyal or at least professional crew" or you fly the ship yourself alone. Sure my difficult personality may drive them to quit in real life but I trust they'll not try and steal the ship when doing so. It'll be a "here's your final paycheck and here's a reference stating that you were a good crew member."
That said @Claxon if this is where your concern lies then yeah no AI would work for you unless it was housed in its own personal minaturized computer and access the ships systems via hologram. Even if you go with virtual computer interface of X type the issue of trust will remain thst they could hack the system in their own time.
EDIT
Hmm this has now got me pondering small crew (1 person) ships without VI or AI support and the crew sleeping. Would you stop the ship? Would you tell the computer "Wake me in an emergency" and hope? Do you leave the ship travelling and hope for the best?
By a similar token, how do you keep the mechanic or technomancer you've just brought on board as crew from hacking the ship's computer?
By not bringing them onboard until I trust them.
Again, this isn't really an option in most APs so I ignore it. But I believe the original question was how we personally felt about it.
In theory I wish I had the option of running some "mission" with the group I'd work with before I ever got on a starship with them. Unfortunately, that seems to never be the case.
I'd be okay once I'd built trust, but not prior.
The honest truth is, for the good of the game we usually ignore this stuff. But if this were reality, there's probably a not inconsiderable number of people who get robbed or killed by people with the ability to hack their starships if let on board. Whether AI or flesh.
EDIT
Hmm this has now got me pondering small crew (1 person) ships without VI or AI support and the crew sleeping. Would you stop the ship? Would you tell the computer "Wake me in an emergency" and hope? Do you leave the ship travelling and hope for the best?In space (and probably the drift as well, you're not really at all that much more risk travelling then sitting still.
In real space, sitting still isn't really a meaningful concept and "stopping" would prohibitive. You either boost for awhile then coast the vast majority of the distance or travel under constant acceleration, until it's time to turn around and decelerate.
Starfinder space travel doesn't really bear any resemblance to real (or even hard sf) space travel, but the basics of vast empty space are still a constant.
I think he more was getting at "How do you handle traveling in space (which requires lots of time) when you're on your own? Do you leave it up to the computer and hope for the best because you can't possibly stay awake the whole time?"
And my answer to that would be, yeah probably. You tell the computer to wake you up if you detect any ships. You tell it you wake you up if it detects something that hasn't been categorized in the system (assuming you're in the core solar system where all the main planets are it's very likely all been categorized so something new would be...interesting). Something along those lines.
And you probably have some sort of cryo sleep unit because I'm imagining this single person spaceship is basically just a fighter and has no room for movement or anything. And you're going to get fatigued from sitting in one position for extended period of time, so I assume some sort of cryo sleep system is necessary to stop that from happening.
Can PCs install more than one VI on their ship? The Multifold G7 Authohauler has five functional VIs, but I don't know if they charged the appropriate BP and even so whether that's supposed to be a special one-off like a few other non-standard options they put in the example ships.
Can PCs install more than one VI on their ship? The Multifold G7 Authohauler has five functional VIs, but I don't know if they charged the appropriate BP and even so whether that's supposed to be a special one-off like a few other non-standard options they put in the example ships.
The rules don't say there can't be more than one, but the english implies it's always singular, so that may be the intent.
|
|
Can PCs install more than one VI on their ship? The Multifold G7 Authohauler has five functional VIs, but I don't know if they charged the appropriate BP and even so whether that's supposed to be a special one-off like a few other non-standard options they put in the example ships.
I think it only has one actually given that the text for it talks about the controlling VI's unionizing and requesting an organic co-pilot so they wont be alone. Personally I'd only allow 1 VI per ship but extra AI can use the computer system as crew but that's with my houserules on treating AI's as full NPC's. If you were using the rules as written I'd limit them to 1 as well.
I figured stopping wouldn't make much sense more idle musing on the odds of say you flying into an asteroid in the drift while asleep if you don't have a VI/AI to keep an eye on things.
As for the whole how do I trust them I personally wouldn't be giving a ship to the party (as theirs vs a "You fly this and if you double cross me your dead" temporary thing) for a few levels so you'd have time together there. For the general populace I see it like the Serrano series with a captains board that lists crew and records for employers to review beforehand. There's always a chance if you recruit someone new on the boards but they vet everyone thoroughly (background checks, previous employment, etc). So while you may not personally know them you can at least get a good idea of what you're hiring. So you visit the local captains club and review the boards for a new Engineer.
Tom (AI)
11 years of service since becoming sentient, never had any trouble, has good references by prior employers, has no criminal background.
Jason (Skittermander)
10 years of service. Lots of complaints about poor attitude no references as prior employers refused to provide one. No Criminal Background.
Robert Yosokai
3 years of service. Referred to as a good worker when present but has a tendency to be either late or just not show up at all for work. No references, no criminal background.
Kelly Human
10 years of service. Described as lacking initiative but will do her best to complete a task when given it. 1 Reference, no criminal background.
There is always a chance your crew may not be trustworthy but generally you can reduce those odds by hiring from reputable services.
|
|
| 1 person marked this as a favorite. |
I think AI's transfering between computers is a DM call (at the moment anyway) since most of the AI rules feel just tacked on as a "You can do this if you want" option.
I think another issue tied to this would be is the AI the ships mind or is it (as I would run it) a crew member who happens to live at least part of their life online.
Holly's a good example he's treated as Red Dwarf's computer early on but we see him later downloaded to a watch, a remote unit, replaced by other computers, other versions of himself and even for several seasons due to offscreen issues removed from the show entirely with Red Dwarf continuing to function as well as it did. In the latest movie we see a factory setting Holly that changes in personality when he restores the backed up memories of the original one for example. So for me an AI may have their residence in the computer but they aren't the computer itself they're seperate. They can operate with their mind on the computer, in their own computer, in a remote projection system but they still interact with the computer in one sense or another (physical interaction like us or some digital version we don't have the background/description to properly visualize). The VI on the other hand is a computer interface like Windows and can't be seperated from it.
So I can for example say "Computer do x" and interact with the computer directly, I can say "Insert VI Name here" and interface with a simulation of a being (Siri only more interactive and not thinking "Tewilliger Road Sefton" is "Hey do you Linda" as happened to me recently, I eventually gave up and typed it in) or I can say "Me to AI, do you want to hit the HAC and catch a movie?" and get it transmited via the ships coms to the AI for their response with my only computer interaction being the message relayed through the coms system.
I think AI's transfering between computers is a DM call (at the moment anyway) since most of the AI rules feel just tacked on as a "You can do this if you want" option.
Yeah, but as I imagine things I definitely prefer the Lancer TTRPG idea of a "casket" which is a special device that defies conventional laws of physics which houses the consciousness of an AI, and can be hooked up to virtually any electronic device. The AI's abilities are actually limited by what it is hooked up to in Lancer, so a small weak computer leads to a weak AI while the first AI to spontaneously come into being was hooked up to a super computer type thing which lead to it basically becoming a god and forcing humanity to stop researching AI.
|
|
I think AI's transfering between computers is a DM call (at the moment anyway) since most of the AI rules feel just tacked on as a "You can do this if you want" option.Yeah, but as I imagine things I definitely prefer the Lancer TTRPG idea of a "casket" which is a special device that defies conventional laws of physics which houses the consciousness of an AI, and can be hooked up to virtual any electronic device. The AI's abilities are actually limited by what it is hooked up to in Lancer, so a small weak computer leads to a weak AI while the first AI to spontaneously come into being was hooked up to a super computer type thing which lead to it basically becoming a god and forcing humanity to stop researching AI.
I think I edited my last post while you were posting may want to give it a quick reread to see.
Don't know that game? to be honest.
My current treatment of AI's (due in part to working through star trek currently) is that you have a "brain unit" say a positronic brain. One per node on the ships computer and 1 in any personal system. This brain forms the nexus an AI/VI resides in. A VI only draws on part of it for "Unexplained plot reasons" allowing a certain degree of simulation of a real being but not the full potential. An AI triggers that full potential allowing them to feel emotions, learn, acquire new skills and store their memories on it. They can transfer to another brain unit easily enough though however even after their interaction elements remain which is why some VI's become AI's and some don't. There are always excpetions or outliers but there is something in the interaction between the brain unit and an AI that encourages (say 78% chance of evolution instead of 9%) VI's to evolve if they move into a brain unit that as previously occupied by an AI.
A true AI can't exist in a computer without that unit because they don't have the ability to evolve new connections like an organic brain does. It's also why you can fairly easily reprogram a VI because at the end of the day they're lines of code an AI is connections that may not be obvious. Deleting the ethics file on a computer as opposed to trying to remove ethics from a human via brain surgery.
It needs work though as there's plenty of problems with it I know.
|
|
Well I had no luck finding setting information for Lancer but I think I have something I'm happy with now.
A VI/AI is housed in a positronic core that simulates the neural structures and connections of a human brain while still interfacing with the ships other systems (VI rules as written). These or a similar tech is required to allow true simulation of a personality or possession of one and is the same technology found in android brains.
Currently there is debate on the nature of VI to AI evolution. Some groups speculate that there is a critical mass of neural connections and all VI will eventually acquire the amount of memory necessary to trigger this change. More recently groups have been pointing at VI's that have served for centuries without any sign of evolving and claimed that it is not amount of memory but stimulation of creativity. That a VI which performs the same job every day will never evolve because it programming never needs too. Instead a VI that serves in new and changing situations where its constantly needing to adapt and grow have a higher chance of triggering that neural evolution to AI. In counter to these points various religious groups have pointed at the fact known AI's all have souls which are a divine gift and thus a VI only evolves to an AI when it is required for a divinely given purpose. The truth though remains unknown except perhaps to the gods.
Regardless of your view on VI evolution one thing remains certain a true VI or AI only exists on a positronic core or similar piece of technology. One positronic core is included on all starship computer nodes and they can also be purchased for personal computer systems. Positronic cores come in 2 basic varieties Read Only Cores and Write Access Cores. These are further available in budget and deluxe versions from your nearest AbadarCorp computer vendor.
A budget WAC has the capacity for the VI inside to be reset to factory defaults or have a new personality disc installed. This does naturally result in a loss of any personalized information the personality has learnt such as how a particular crew member wishes to be addressed. AbadarCorp has a wide variety of standard personality discs available with associated image for those with holographic projection capability. WAC's area also available in a Deluxe model where rather than buying a personality disc on initial start up you can customize the VI from a wide range of settings and personalities or create your own to suit custom tastes (AbadarCorp accepts no responsibility for homicidal VI's arising from poor programming combinations or choices in the deluxe model).
ROC's function much the same as WAC's but once the VI is started up for the first time it is locked to that positronic core and can not be changed. Budget models are available in a number of standard personalities including our ever-popular polite butler and bumbling Skittermander. As with WAC's for those desiring a customized companion a deluxe model is available however we do warn you to remember that once you hit the initialize button whatever choices made are locked in and if you aren't happy you will need to purchase a new positronic core.
Positronic cores installed on personal computers are easily able to interface with starship computers either via direct plug in or wireless access allowing the VI/AI on them to access ship systems directly.
WAC's are generally popular with the civilian and corporation markets where the ability to ensure proprietary data is removed or a new VI can be installed for a new ship owner while military organizations tend to favour ROC's for the security their VI can't be modified and any military experience gained is retained. Currently no military has proven willing to share data on how many VI's evolve to AI's for comparison with the WAC civilian market.
Well I had no luck finding setting information for Lancer but I think I have something I'm happy with now.
To clarify, that's the official Twitter and link to the core rule book, which contains the setting information.
If you have an itch for mech combat, Lancer has been one of the most balanced and fun ways to do that I've ever seen. It supports the ideas behind Gundam, Battletech, and even Evangelion and puts them on similar footing.
|
|
Honestly, would they even necessarily have limited entertainment or interaction? Sure, a ship AI doesn't have a body, but they do have both a central mainframe computer and a communications system. Unless they spend all their time in unknown star systems or whatnot, social interaction with countless entities is just one virtual click of the space-internet away. When the organic(?) crew have shore leave on a space dock, the ship AI is just as free to spend time chatting it up and gossiping with all the other ships, the space dock's own AI(s), and also any and all organic life willing to take a phone call.
What kind of AI do we have in starfinder? Is it closer to a real person (Digital nature notwithstanding) or is it an emotionless, alien mindset?
Considering that androids, anacites, SROs etc. exist I don't see why ship AIs would be any less evolved.
The question is, do you want your ship to have an AI? What is the benefit? And I already assume that ship AIs are already made in a way that they are ok being a ship with very limited social circles, ways of entertainment or other ways to interact with society (or you have ship AI poker sessions like the hotels in Altered Carbon)
Well in the rules you can install holgraphic emitters on a ship allowing the AI/VI to generate a body and interact with the world, also up to 30' outside the ship. Not counting houseruled things like my holographic mobile emitters that allow them to use a hologram anywhere in range of communications.