4
$\begingroup$

This question already has an answer here:

I was thinking why there would be crew on a space warships when everything can be automated by even basic AI.

Without crew you can pull all the gees you want with your torch-drive and no need for artificial gravity centrifuges. I thought about maybe dozens of auto-warship frigates and destroyers around a single flagship with crew controlling the different groups of warships meanwhile all the ships provide protection and point defense for the flagship.

So is there really a need for crew on space warships as depicted in many military-science fiction stories? Just slap a fusion drive down with a bank of thermonuclear warheads completely covered by CIWS and laser point defense and coil guns.

When the light-lag gets too long just let the AI take over, it'll probably even do better than the flagship crew anyways. Maybe the warships can even deploy mirrors to protect the flagship from the incoming laser fire. I can't find a real reason for crew apart for maybe repair and just in general story telling. I mean it does provide for some awesome war stories.

$\endgroup$

marked as duplicate by Mołot, Mark Olson, Logan R. Kearsley, Renan, Anketam May 29 '18 at 22:43

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • 1
    $\begingroup$ This seems very close to worldbuilding.stackexchange.com/q/113502/627. $\endgroup$ – HDE 226868 May 29 '18 at 19:14
  • $\begingroup$ This seems precisely the point made by Iain Banks' Culture Novels $\endgroup$ – user535733 May 29 '18 at 19:31
  • 1
    $\begingroup$ sure, if drama counts as a reason $\endgroup$ – jean May 29 '18 at 20:30
  • 2
    $\begingroup$ NOTE TO VTCers. I completely disagree that this is a duplicate of the stated question, which specifically asks why live crews would be used to fight wars. There is a subtle but important difference between these questions. @Mołot, if you disagree with me, please post your reasons. Thanks! $\endgroup$ – JBH May 29 '18 at 21:11
  • 2
    $\begingroup$ This is a very important question I want answered as well, ans as JBH says this is an importantly different question compared to the others. I personally only have two not terribly good reasons: AI including weak AI cant be trusted (and non-AI programming is too easily countered), or theres a mcguffin in play that prevent high levels of computer power (I like introducing high fluctuations that mess with small electronics so you return to larger cirquits offering less modern era computerpower that suit your setting). $\endgroup$ – Demigan May 29 '18 at 21:59
3
$\begingroup$

The main issue about a robotic fighting ship is control. If it isn't smart enough to act on it's own, you need to be close enough that the light speed lag won't affect fighting performance. If it is too smart, how do you control it?

Not Smart:

We had fly by wire missiles and torpedoes. They would be the beginning of this. You had to stay within "wire" distance of the target.

then we had radio controlled weapons (missiles, torpedoes, drones). The problem with this tech is distance (signal strength and time delay) and interference. Theoretically, an enemy could intercept communication and/or give false commands to the weapon.

To get around this, you need to program in just enough smarts to go after the target once designated. We kind of have this now. We have missiles that, once targeted, follow or even switch targets on their own. You have to either be close enough to set the target or give it enough smarts to be able to select it's own targets. However, this starts to get into the next category.

Smart:

If the weapon can select its own targets, you better be sure that it selects the correct targets.

If it selects a non-enemy target, or worse, gets captured and turned on you, you will not be having a good day.

I suggest you read Fred Saberhagen's Berserker series. It is about self replicating AI war machines that outlast their makers and continue their war.

The Exterminator movie has a similar take on this theme.

There are a lot of stories in which smart weapons go wrong. Some because of buggy software, some with battle damage and others where they were reprogrammed. Unless you want to repeat those stories, you better make sure that you take them into account to show how that didn't happen.

$\endgroup$
  • $\begingroup$ Thank you for pointing this out for me. Helps me draw the line between the types of AI on such warships. I was thinking more of this type of AI: if enemy detected then plot course to intercept and destroy, if unable, contact nearest human operator, if not possible: defend to best ability. (But at the same time this AI would not be very smart and could not compensate, be creative, etc.) $\endgroup$ – zertofi May 29 '18 at 20:01
  • $\begingroup$ Here's a problem for you to solve (or weave into your story): Can compensate leads to self programming (that can go all kinds of wrong). Can't compensate means that the enemy may figure out a way to outsmart all of the weapons. $\endgroup$ – ShadoCat May 29 '18 at 22:02
  • $\begingroup$ That's important too. What if it is AI vs AI? Like I mentioned in the comment under this answer . Here there are no humans to fight anymore, similarly this happens at a point in my story resulting in auto defense and attack systems fighting each other over the remains. $\endgroup$ – zertofi May 29 '18 at 22:09
  • $\begingroup$ It could happen for a short while. However, think of how much infrastructure it takes to support one fighting craft. Is it all automated? How mines the minerals, then smelts them and then builds the parts that then need to be assembled and then delivered? $\endgroup$ – ShadoCat May 29 '18 at 22:21
  • $\begingroup$ Nope, the infrastructure is upheld by human coordination. Sure rearming and refuel is automated but that could only go on until the stations run out of ammo and fuel. It is more of a cold-war gone nuclear scenario. Every automated weapon is launched if humans are compromised. So every orbital thermonuclear platform and every interplanetary destroyer goes ahead and fires its payload into the enemy until both sides run out of ammo or every planet in the system is so irradiated that nothing works anymore and every combat unit is dead. Kind of the moral I try to convey about current automation. $\endgroup$ – zertofi May 29 '18 at 22:31
2
$\begingroup$

Expense is one reason. It depends on your world, but in ours, robots with fine motor control are pricey, and human repairmen are much less so. Not to mention the fact that humans can innovate and find ways around the "irreplaceable" part that broke down, while robots can't. (unless they're "truly" intelligent and not merely imitations) A ship fully reliant on AI would also be unlikely to be capable of colonization or transport missions without crippling its capabilities. Even if the pros outweighed the cost in other use-cases, in these cases, (where you have to maintain a livable space for humans, anyway) a human crew is probably much more cost effective without losing much in the way of effectiveness. (especially if they have an AI to assist them)

$\endgroup$
  • 2
    $\begingroup$ See the Isaac Asimov short story “The Feeling of Power”, in which a society that has only AI-controlled weapons wins a war by re-inventing human-piloted weapons, which are much cheaper and easier to produce. $\endgroup$ – Mike Scott May 29 '18 at 19:21
  • $\begingroup$ @MikeScott That does provide for some ethical and moral points of war that you could cover in a story which can show how it effects society and people. Very interesting. $\endgroup$ – zertofi May 29 '18 at 19:31
1
$\begingroup$

If you can't have strong A.I., and we have no reason to believe you can at this point in time, then tasks that require intuition rather than procedural logic require humans not machines. Also if you reverse this slightly if you can have strong A.I. it's probably too smart to go to war, war is not logical but you can get humans to fight if you tell them it's "right" enough times. Either way you need emotional, biological beings if you want to go to war.

$\endgroup$
  • $\begingroup$ That makes sense. Would the idea with the flagship work? Maybe smaller flagships controlling small fleets of warships which split from the main fleet? $\endgroup$ – zertofi May 29 '18 at 19:12
  • 1
    $\begingroup$ @zertofi Only if you have instantaneous communication over large distances (those measured in things like light-minutes), otherwise the light-lag kills you, fast. $\endgroup$ – Ash May 29 '18 at 19:14
1
$\begingroup$

There’s no particular reason to believe that an AI would make a good warship crew. It would be walking a very fine line between being smart enough to be as effective as a human crew, and being smart enough to immediately desert. Or the three laws of robotics may apply, or the society may value AIs equally with humans (e.g Iain M Banks’s Culture) and not want to put them in harm’s way.

$\endgroup$
  • $\begingroup$ What if the AI's sole purpose is to attack and defend against the enemy? It only has the intelligent capacity to figure out were to fire and how to intercept the enemy, nothing else. Would't this be more effective than crew? Kinda like assembly line robots, they only do the task at hand because that is what they are programmed to do. $\endgroup$ – zertofi May 29 '18 at 19:17
  • 1
    $\begingroup$ @zertofi Such an AI likely wouldn’t be smart enough to innovate or respond to the unexpected, so it would need a human crew for that. $\endgroup$ – Mike Scott May 29 '18 at 19:18
  • $\begingroup$ If the AI resembles current technology in any way, it's easy to set the AI's goals in such a way that it would not "want" to desert. Current AI (not sci-fi it's-a-robot-with-a-soul AI) only "wants" what it's been programmed to "want." $\endgroup$ – Josh May 29 '18 at 19:21
  • $\begingroup$ I see, so you could have a central control point in a fleet like the flagship I described. Things like point-defense is handled by AI meanwhile the tactics themselves are handled by the crew. Maybe there are multiple central warships with crew with dozens of auto-warships. $\endgroup$ – zertofi May 29 '18 at 19:23
  • 1
    $\begingroup$ @Josh But if the AI resembles current technology in any way, then it’s far too stupid and predictable to win a battle. We’re talking about AIs that are many orders of magnitude smarter than anything we have now. $\endgroup$ – Mike Scott May 29 '18 at 19:23
0
$\begingroup$

there are concurring factors:

  • cost: an AI costs quite some money, you don't want the enemy to hack it undisturbed or, worse, that a BSOD or a critical patch install makes it useless in the middle of the battlefield
  • lack of trust: do you trust that the AI will be always patriotic and loyal, 100% hacking proof?
  • tactical improvisation: an algorithm can be cracked and its pattern discovered and neutralized. Creative thinking not.
$\endgroup$
  • 1
    $\begingroup$ "an algorithm can be cracked and its pattern discovered and neutralized. Creative thinking not." Damn that's a cool sounding phrase, mind of I use it in my writing? $\endgroup$ – zertofi May 29 '18 at 19:34
  • 1
    $\begingroup$ @zertofi, you can use it. I got inspiration from dilbert.com/strip/2012-03-18 $\endgroup$ – L.Dutch - Reinstate Monica May 29 '18 at 19:35
  • 2
    $\begingroup$ Psychology can be applied to dictate the outcome. No matter how much "creative thinking" you throw at it, the machine can look at your options, analyse your flight patterns from approach and onwards and predict the likeliest outcome(s). The best (worst) part is that the AI can predict all outcomes that will mean something and deal with those causing a no-win scenario. Creativity would only mean you picked an outcome with a low prediction and is at best a stay of execution. $\endgroup$ – Demigan May 29 '18 at 22:09

Not the answer you're looking for? Browse other questions tagged or ask your own question.