[Space] Autopilots and Autoturrets

[Gearloose]

I'm planning a hard sci-fi campaign featuring space travel. While reading the systems for space travel and combat in Spaceships and Interstellar Wars, I've realized. Both systems assume the ships are operated manually, like in Star Wars, with characters manning gunner stations and hotshot pilots at the helm.

While this fits perfectly a space opera game, I'd like to have my game with more automated ships, like real-life spacecraft are.

So, what you guys think would be the best way to approach this? Just give the shipboard AI Piloting and Gunner skills? Or is there a system somewhere to command spaceships at a "higher level" than manually piloting it?

[Ulzgoroth]

So the easy, GURPS-y solution is to keep those things operated by characters...but have those characters be AIs, potentially quite basic AIs. One difficulty is that there's no general guidance on limiting AI skill levels - assigning numbers that work for your setting is an exercise for the GM with no textual assistance.

In some cases, you might want to render more basic automation as not even an AI, but a program that does very specific things with no higher mentality or sense of self at all. It might still have specific skill levels, if it does things that require rolls. This might be appropriate for modern airliner autopilots, space probes executing programmed burns, gunlaying software, or automatic point defense systems. In many cases, you may want to roll for the person who gives the software its instructions - programs like this do what they're told and often don't have sufficient checks to make sure that what they're told is at all sensible.

[AlexanderHowl]

The problem with using AIs in a story is that it removes the human angle. In addition, AIs tend to be quite dumb in Spaceships. For example, at TL10, a SM+4 or SM+5 spacecraft can only support one IQ 10 NAI and you need to be SM+8 or SM+9 before you can support one IQ 10 SAI. Thanks to TL10 genetic engineering though, the average human can have IQ 12.

Now, how does Complexity relate to DX? I tend to give AIs a DX equal to their IQ, as physical activities are just as difficult as mental activities in GURPS, but everyone has their own take. In general though, I think that it would work for Spaceships as a good rule.

While a TL10 human will only have an average of DX 11, a military can just recruit above average people. If the average gunner has a DX 13, they will outperform any SAI of less than Complexity 10 (SM+12 or SM+13), and a spacecraft can probably have dozens or hundreds such gunners. In any case, a government which depends of AIs will find itself outperformed in combat by a government that uses humans, at least until TL11.

[Ulzgoroth]

Originally Posted by AlexanderHowl

The problem with using AIs in a story is that it removes the human angle. In addition, AIs tend to be quite dumb in Spaceships. For example, at TL10, a SM+4 or SM+5 spacecraft can only support one IQ 10 NAI and you need to be SM+8 or SM+9 before you can support one IQ 10 SAI. Thanks to TL10 genetic engineering though, the average human can have IQ 12.

You realize you can put a computer on a spaceship, right? The integrated computer networks have lousy complexity, but a microframe is smaller than a crew member and a macroframe is less than half the mass of a cabin.

Originally Posted by AlexanderHowl

Now, how does Complexity relate to DX? I tend to give AIs a DX equal to their IQ, as physical activities are just as difficult as mental activities in GURPS, but everyone has their own take. In general though, I think that it would work for Spaceships as a good rule.

There is no rule connecting them. Your take isn't unfair - it's not too implausible for full-up dexterity to even be harder for computers than full-up IQ, though we can't really tell since they don't have either thus far.

OTOH that's pretty irrelevant here since the AIs don't need high stats - they need high skills.

[Agemegos]

Whether automatics are better than humans, about human maximum, comparable to skilled humans, comparable to humans with basic skills, or poor is essentially a setting switch. Remember that the same technology is probably in use in agriculture, mining, manufacture, manual services, cognitive services. Dial it high if you want a post-labour, post-industrial future.

So far as I know there are no rules or guidelines, either for the abilities or the cost of sensors, processors, and actuators.

[Ulzgoroth]

Originally Posted by Agemegos

Whether automatics are better than humans, about human maximum, comparable to skilled humans, comparable to humans with basic skills, or poor is essentially a setting switch. Remember that the same technology is probably in use in agriculture, mining, manufacture, manual services, cognitive services. Dial it high if you want a post-labour, post-industrial future.

Er, it's pretty easy to have automatics be very good at, say, controlling a spaceship or a weapon turret, and really bad at mining, agriculture, and most other forms of blue or white collar work.

I mean, that's not even speculative fiction.

Originally Posted by Agemegos

So far as I know there are no rules or guidelines, either for the abilities or the cost of sensors, processors, and actuators.

Uh, there are quite a bit of RAW on cost of sensors and processors. Actuators are irrelevant unless you're doing something unusually low tech - human operators are already going to be using computer interfaces anyway.

[Agemegos]

Originally Posted by Agemegos

So far as I know there are no rules or guidelines, either for the abilities or the cost of sensors, processors, and actuators.

And servos.

[Agemegos]

Originally Posted by Ulzgoroth

Er, it's pretty easy to have automatics be very good at, say, controlling a spaceship or a weapon turret, and really bad at mining, agriculture, and most other forms of blue or white collar work.

I mean, that's not even speculative fiction.

Barely. On current trends we'll have automatic farms and mines before we have armed spacecraft. Research assistants, juniors in law firms, junior programmers, and physicians (internists) are on the skids already.

Uh, there are quite a bit of RAW on cost of sensors and processors. Actuators are irrelevant unless you're doing something unusually low tech - human operators are already going to be using computer interfaces anyway.

Human operators will be controlling everything by servos anyway, so everything should be assumed to include servos sufficiently strong and deft to effect the best possible human skill?

[Ulzgoroth]

Originally Posted by Agemegos

Barely. On current trends we'll have automatic farms and mines before we have armed spacecraft. Research assistants, juniors in law firms, junior programmers, and physicians (internists) are on the skids already.

I don't think there are any current trends promising fully automatic farms or mines. While some parts of farming are highly mechanized, others are strong candidates for the least-mechanized part of the modern economy, and mechanizing them would require better-than-cutting-edge robotics to replace extremely low paid human labor. A robot farm for certain crops might not be too far off from achievability, but a fully automated agricultural sector is not at all close.

Meanwhile, the lack of armed spacecraft for them to operate has no bearing on the ability of computers to perform the tasks in question.

Originally Posted by Agemegos

Human operators will be controlling everything by servos anyway, so everything should be assumed to include servos sufficiently strong and deft to effect the best possible human skill?

If that's how you want to put it.

[AlexanderHowl]

Would you really want an automated warship though? The great thing about a human operated warships is that it is very hard to suborn the ship unless the majority of the crew engages in mutiny. Conversely, you only have to compromise the central computer on an automated warship to suborn it.

When it comes to SM+4 to SM+6 fighters, automation is an acceptable risk because it is not really that much more likely than a mutiny. As the warship gets bigger, however, the probability of a mutiny decreases, so the relative risk of compromising the central computer increases. When it comes to SM+12 or larger capital ships, it is an unacceptable risk because the chance of mutiny ends up being much less than the chance of an automated warship being compromised.

In my mind, a rational space military possesses manned capital ships supported by automated drones. While massive numbers of drones protecting against kinetic weapons, large beam weapons end up being the weapon of choice. At TL10, that would mean lasers (UV or normal) for long range sniping or particle beams for close brawls.