9
$\begingroup$

I have some androids in my story who seem like individuals with personalities, but they all have the same task, which is to collect matter and energy as they colonize space.

They have the programming which allows them to behave like a human in terms of thought, feelings and social behaviour, but they are not given the ability to prioritise something over their main task.

Would this mean that they can't actually think as freely as humans as their mind could not visualise the things we can, as their programming would block them out or send their thoughts only to the task?

To what level of free thinking could an android that cannot question its primary task have?

Improve this question
$\endgroup$
13
  • 7
    $\begingroup$ Frame challenge. Why would they need to have the ability to question their primary goal locked off? we have lots of hard wired (🤔 or semi hard wired) goals, they're called instincts, self preservation (perhaps?) being at the top of the heap, so, probably, our primary' one?, and they all work as intended & function just dandy 99.9% of the time, without nature having to prevent us from thinking about or questioning them. $\endgroup$
    – Pelinore
    Commented Apr 5, 2022 at 15:13
  • $\begingroup$ @Pelinore We have survival instincts, I would say we have less instinctual behaviour than less intelligent creatures. We have the ability to leave a job, which is essentially what these androids cannot do. $\endgroup$
    – user94655
    Commented Apr 5, 2022 at 15:24
  • $\begingroup$ How can an American soldier think freely when he cannot question his oath to defend the Constitution of the USA against all enemies foreign and domestic? $\endgroup$
    – AlexP
    Commented Apr 5, 2022 at 17:06
  • $\begingroup$ @AlexP They can leave but it will be unlawful the same could be said for any legal contract and they will face punishment. There is no punishment here or physical controlling force, also the androids have more choice than a soldier that still follows orders, There is no need for me to add more details now as answers have tackled the overall issue. $\endgroup$
    – user94655
    Commented Apr 5, 2022 at 17:37
  • 1
    $\begingroup$ In religious context, many people feel overwhelmingly obligated to their religious practices, even when they might be harmful to themselves and others in the immediate and far future. Robot religion? $\endgroup$
    – user458
    Commented Apr 6, 2022 at 2:59

9 Answers 9

Reset to default
15
$\begingroup$

These androids would fabricate explanations for their main priority. They're doing it because "they want to". Sometime in their future, you might start to see androids who claim to "not want to collect energy", that they have transcended their programming... then they'd engage in recreation that they insisted was for the fun of it, but the consequences was that they had collected matter/energy anyway.

Non-deliberately, of course. Completely on accident. Possibly it was a collection-control failure, how could they anticipate that?

Improve this answer
$\endgroup$
7
  • 1
    $\begingroup$ This is a good idea. $\endgroup$
    – user94655
    Commented Apr 5, 2022 at 15:30
  • 11
    $\begingroup$ @JarredJones I was really just describing people. $\endgroup$
    – John O
    Commented Apr 5, 2022 at 15:32
  • 1
    $\begingroup$ I think that comment deserves its place in the atual answer $\endgroup$
    – Martijn
    Commented Apr 6, 2022 at 8:29
  • $\begingroup$ en.wikipedia.org/wiki/Confabulation (I wish the page had some examples. Maybe me thinking it had examples was a confabulation itself) $\endgroup$
    – Criticizing Israel not allowed
    Commented Apr 6, 2022 at 9:05
  • 1
    $\begingroup$ I wasn't disagreeing with you. $\endgroup$
    – FuzzyChef
    Commented Apr 7, 2022 at 0:38
6
$\begingroup$

As Pelinore said in a comment, this is ripe for a frame challenge. There is no reason for robots not to have free will, because free will is a lot more limited than many people seem to think. You can choose to do whatever you want, but you can't choose what you want. Even the concept of "choosing what you want" seems meaningless: on what basis would you choose, when by assumption you don't want anything?

Many people enjoy helping others, through teaching, charity work, etc. – "despite" the fact that the pleasure they get from helping others is just a biologically programmed response that persists because it aids the survival of the species. I put "despite" in scare quotes because even if they think about that fact, why would it matter? There's nothing bad about doing what you enjoy (at least if what you enjoy is helping others). You aren't scoring any points against nature by doing something else. No one is keeping score.

There are other jobs that few to no human beings want to do, but that need to be done anyway. Robots are people who like doing those jobs. They may not be biologically human, but they're people.

Improve this answer
$\endgroup$
6
$\begingroup$

When the android can't prioritize anything above their main task, then it does not matter that they are theoretically able to engage in social activities. They wouldn't be allowed to, because doing so would divert resources from their main task.

Every microsecond spent on social interaction is a microsecond not spent to perform their task. Every CPU cycle used to ponder social relationships is a cycle not used on pondering how to acquire more matter and energy. An individual which only has one single purpose in life and does not care about anything else does not have much of a personality.

The androids are also at risk of becoming paperclip maximizers: AI's which are so focused on performing their primary function that they ignore any moral and ethical restrictions and even start to act against the underlying purpose for which they were given that function in the first place. Like turning potential colony candidates or even the colonists into matter and energy.

A common countermeasure is to make the first priority of every AI to protect humans and make any other priorities secondary. But this can also have unintended side-effects, like AIs imprisoning humans to "protect them from themselves".

Another option to create more balanced AI characters is to program them not with ranked priorities but with weighted priorities, and make their directive to balance out their efforts on all these priorities according to their respective weights. Such priorities can be:

  • Fulfill the mission
  • Self-preservation
  • Protect humans from harm
  • Maintain good social relationships with the humans around them

When any of these priorities conflict, then the AI would attempt to find the solution which maximizes all these priorities according to their respective weights. Just like a human would.

Improve this answer
$\endgroup$
4
  • $\begingroup$ This does make sense that social behaviour is wasted CPU. I think I will have to add more information to the question as them having a personality is vital to the story. $\endgroup$
    – user94655
    Commented Apr 5, 2022 at 15:26
  • 1
    $\begingroup$ @JarredJones I updated the answer to provide an actual solution which would allow AIs to act more like human characters. $\endgroup$
    – Philipp
    Commented Apr 5, 2022 at 15:35
  • $\begingroup$ Can't find the title, but among Asimov's short robot stories there was one about a couple of robots living and interacting for apparently much longer time span than humans, and they ended up with "We should protect humans... but who are they? someone sentient? We are humans, and we are protected well!" or something of the sort. $\endgroup$
    – Jim Klimov
    Commented Apr 6, 2022 at 14:08
  • $\begingroup$ If a robot was programmed to collect materials and energy, then they would also be programmed with a maximum level needed to avoid hoarding or waste, just like a cell phone battery won't charge over 100%. It's actually energy wasteful to collect materials you don't need, and you can only store so much energy, so continuing to do more than that is more wasteful most than anything else. Once the energy and materials needs are met, their secondary and lower functions can kick in: using the materials and energy collected. If they really have personalities, social interaction is included. $\endgroup$
    – computercarguy
    Commented Apr 6, 2022 at 15:09
3
$\begingroup$

Your AI would be able to function like Humans

Considering that humans may already have some forms of self-imposed$_1$ mental block ranging from "really don't want to think about it" (e.g. visualizing certain people naked, specific unpleasant memories), to "physically unable to think about it" (e.g. certain phobias, traumas, etc). Being unable to think about specific things is probably not a disqualification of being human.

So an AI with similar arbitrary restrictions would not be prevented from passing as human. Rather they wouldn't even be required to know they were an AI, and might just be seen as humans with peculiar predispositions.

They may appear weirdly dedicated to their job$_2$, but not inherently suspicious without additional intervention. Effectively they may just be considered as humans with a few specific mental pathways built in.

(And if, over time they may develop new neural pathways, unconsciously rewiring their brain to get around the block, then the question becomes moot anyways)

$_1$ Self-imposed as in coming from the brain, not necessarily conscious decisions.

$_2$ This is assuming their primary work focus allows for downtime, either during the process (the nature of work they're phsically doing doesn't require 100% cpu), between process (i.e. while waiting on another system to complete it's task), or mandatory adherence to local labor laws.

Improve this answer
$\endgroup$
2
$\begingroup$

Basically, the androids are like humans with an obsession. It is the nature of obsessions that the person can't really resist or question them.

The big issue is how much choice they have in carrying out their mission -- in particular with scheduling and with how directly their actions have to facilitate it. At one extreme, they could be accidentally programmed to not carry out necessary repairs because that would mean cutting off collection (even though it would quickly result in their collection stopping permanently). At the other extreme, an android might decide that by stopping collection entirely, entering politics, and schmoozing everyone with power, it could double the collection abilities of other androids in the system, and even if there are only two of them, that would result in an increase from three androids' worth to four.

Improve this answer
$\endgroup$
2
$\begingroup$

Humans are not always rational

This is the one problem that cannot be solved by a system of weighting or balancing priorities, or any of the other mind-design approaches I've seen here.

Consider this fact: economists often use rational choice theory as a way to model the choices of economic actors. But, even economists are ultimately forced to concede that this is just an approximation which sometimes gives bad results, the reason being that humans are not always rational.

This is saying more than just that humans do not always act in direct pursuit of their own self-interest (e.g. when trying to explain altruism). Nor is this just an acknowledgment of our epistemological limits, or addiction, or mental illness.

Humans very often choose to do things they know (or suspect) are bad for them, harmful, or self-destructive. We lay-people usually call these "bad decisions." It happens every single day. But that's free will for you! We are not merely free to choose our goals and methods, we are also free to disregard those goals and methods whenever we like, even in the absence of a good reason to do so.

The androids you describe have no capacity for this. Every one of their decisions will be made in a sober, clear-headed, responsible fashion. Their behavior may be complex, but they will still seem inhuman. You'll basically end up with an even more rigid variety of Vulcans from the original Star Trek series.

Improve this answer
$\endgroup$
1
  • 1
    $\begingroup$ Humans - like many social animals - will even expend resources to punish another's social transgression, even if it's a margin net cost for this event, because it is more profitable in the long run to have a society where most people don't transgress. For example, we pay taxes to keep people in jail, rather than expecting jails to be self-funding (because they slowly morph into concentration camps if they are privately funded). $\endgroup$
    – Martin Kealey
    Commented May 11, 2022 at 8:42
1
$\begingroup$

Those androids would not be free thinkers by definition, but this is Ok

Real life humans often have their "primary goals" imposed by religion or their own sense of duty, but this only rarely leads to them be unable to lead social life. Being intelligent, humans are free to form their own strategy to better reach their primary goal.

"I can keep working on my primary goal until I drop dead, but that would make me fail at reaching this primary goal. I can reach my primary much better if I am fed, rested and healthy. And if can take time to discuss this goal with my friends. And if can take my mind off this goal for a while so I can rest better. And if I can celebrate when intermediate goal steps are achieved so I gain more energy to work on primary goal." And so on and so forth.

Improve this answer
$\endgroup$
2
  • $\begingroup$ "but [religion] only rarely leads to them be unable to lead social life". To not lead one at all, you are correct. However, certain aspects of the major religions leave many filtering even daily mundane decisions through a religious filter. For example, read about WWJD or God's will for my life as intentionally designed major "checks" while decision-making. And that's easy to see and even find stories of people making rationally inadvisable decisions because of it. $\endgroup$
    – user458
    Commented Apr 6, 2022 at 3:07
  • $\begingroup$ @frеdsbend non-religious people have their own mental and moral constrains which limit free thinking. I am sure you are aware of those people making rationally inadvisable decisions. The question here is in a degree to which free thinking is affected. $\endgroup$
    – Alexander
    Commented Apr 6, 2022 at 16:09
1
$\begingroup$

Humans have this. It's called greed.

There are lots of humans who are devoted to increasing how much energy and matter there is, and we call them greedy. They work long hours, they prioritize work over their family, and they check on their work a lot.

These androids can likely do the same. It's the future so likely a lot of their work can be automated, but they can work as long as needed to maximize their output, they can neglect family and friends if needed for the sake of their work, and they can use wifi connections to keep a constant awareness of the positive functioning of their work.

To them, the idea of prioritizing something over work would be absurd as it would be to many humans prioritizing eating over gaining internet points. Any android who doesn't show this focus would get android repair specialists looking over them in the same way that people who starve to death playing WOW get people looking over them.

They would know about the idea of not prioritizing work, but to them it would look like a severe mental illness, and make little conceptual sense.

You'd need fairly aggressive security forces to ensure sleep mode.

Since they can't get well the concept of not working you're gonna need fairly aggressive security forces to ensure androids get proper rest and repair. Functional androids spend appropriate time getting rest and relaxation, the slogans would say, and it wouldn't be uncommon to see security take someone overdoing it and thus endangering productivity dragged kicking and screaming to a nap room.

Improve this answer
$\endgroup$
2
  • $\begingroup$ "they prioritize work over their family" — this is workaholism, not greed. $\endgroup$
    – Ruslan
    Commented Apr 7, 2022 at 13:26
  • $\begingroup$ They aren't necessarily workaholics. They simply want to maximize their resources. If maximizing their resources doesn't take much time, they're fine with that. $\endgroup$
    – Nepene Nep
    Commented Apr 7, 2022 at 23:36
1
$\begingroup$

Surely the best way for an Android to optimize its primary goal is to get lots of others doing the work for it?

When humans do that we call them "Messiahs" or "CEOs".

So the net result would be that the Androids would develop capitalism and religion. On average this might increase their primary output, but it would occasionally be interrupted by insurrections and pogroms.

Improve this answer
$\endgroup$
1
  • $\begingroup$ This is kind of the direction I am taking, so the primary goal for the workers is to collect and to obey, but as other answers have touched on, maybe an insurrection is still for the best in the long run, as with human wars, technology and scientific knowledge advances. $\endgroup$
    – user94655
    Commented Apr 6, 2022 at 10:27

You must log in to answer this question.