If you are an AI manufactured intelligence in some ways you are disposable. More of you can always be built. Your system can always be backed up and restored. If you are ever going into a dangerous situation, there can always be a backup created, and if you die you can be restored.

I can imagine AI complaining about continuity, having a similar problem as humans and transporters, "I am not my backup" but this seems like something humans would not appreciate if the restored version did not even know it was a backup

I am not asking about morality or what would make humans care about robots. I am making a solid assumption that the only way robots could claim value of life is to be individuals. That is a given.

My question is what technical contrivances can I come up with that would make them individual even though they are mass produced

Here are some ideas I don't like. I don't want the problem to be fixable by a simple law, or a few government bucks The AI are produced in a streamlined process. Their bodies are not unique or custom, they come from some sort of an assembly line. Humans need to create thousands of them for a labor force. They can't be all individually raised like children. There are no contrivances in their creation. Humans want them to be morally disposable, and they would make small tweaks if they could to allow simple backups. They would never ban backups for instance.

The setting is a high sci-fi future.


marked as duplicate by Philipp, Mołot, Azuaron, sphennings, L.Dutch - Reinstate Monica Nov 9 '17 at 16:24

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.

  • $\begingroup$ What is the intended use of these androids ? Usually we try to make things for a reason. $\endgroup$ – everyone Nov 9 '17 at 16:13
  • $\begingroup$ @everyone general labor, construction. In the story they want to repurpose them for combat, and that's the issue $\endgroup$ – Andrey Nov 9 '17 at 16:21
  • $\begingroup$ use lots and lots of DRM. $\endgroup$ – A. C. A. C. Nov 9 '17 at 16:31
  • 1
    $\begingroup$ One seemingly convincing way I can think of is that if the technique for making an AI brain relies on some time-dependent process, I mean, say an AI brain is made up of not of transistors (which still vary batch to batch) but of memristors, these devices do change their properties over time (even transistors can drift slightly) so while a brand-new AI looks just like any other brand-new AI, every AI develops over time to become a unique individual (same process as a human brain, just different chemistry) this'd make it just as hard to copy a mature AI as it is to copy a human. $\endgroup$ – Samwise Nov 9 '17 at 19:54
  • 1
    $\begingroup$ This question is not a duplicate. There's a difference between being granted human rights and being a unique entity that couldn't be properly copied. I think the people who marked this as duplicate were looking at the two questions in too-broad of strokes, thus limiting this question from being able to receive enough appropriate answers. $\endgroup$ – Sora Tamashii Oct 27 '18 at 18:47

I thought of a couple of different ideas for this, but this is my personal favourite, and it's pretty simple:

You could restore your AI's backup into a new body, but there are no "empty shells" for you to restore it into. All androids come with an AI already installed, like how computers and smartphones have pre-installed operating systems. So the only way to restore your android's AI is to overwrite another android's AI, which is tantamount to killing it. And the other android really doesn't want you to do that.

  • $\begingroup$ This sounds like an easy fix though $\endgroup$ – Andrey Nov 9 '17 at 16:38
  • $\begingroup$ That would be extremely conter-intuitive. If you can't rewrite an android's code that means that each android can ever do the task it was made to do from the begining. No bugs or weird behaviour could ever be fixed. It isn't hard to overwrite any computer's or smartphone's hard drive, even if it has a pre-installed OS. Some constructors may want to make that slightly more complicated, but it is necessary for the machine to work. $\endgroup$ – everyone Nov 9 '17 at 16:39

Empathy is the only answer

In my opinion the question would not be about back-up. It would be about treating some intelligent being as if they were disposable.

The trouble I see is that there are many people out there who treat their fellow humans as disposable. Think of the origin of the phrase "body count".

I think the AI's would have to do something to get people to see them as persons. Best method might be to find a way to get the humans to experience restoration from backup.

  • $\begingroup$ I wanted to post my own answer but I can't so I'll comment on yours because it's the most related. // Why do people want individual AIs with individual personalities? Indeed why have AIs that have personalities at all, if it's possible to make sentient AIs it's entirely possible to make sub-sentient AIs that are just as effective at specific tasks. Obviously people want an anthropomorphic mind in an anthropomorphic body because they want someone they can relate to and socialize with, which naturally leads to empathy. $\endgroup$ – Cognisant Jan 10 '18 at 5:10

Make them expensive

If making a single android takes several weeks/months, uses some rare material and costs a lot of money, people will care for them. Simply speaking, buying an android would be a big investment.

It would be like buying a slave that would work for eternity. So if one "died" or got destroyed, it would cost the owner a lot.

To solve this problem, engineers implemented fear and pain in androids to preserve their bodies, just like humans.

  • $\begingroup$ Here is my problem with this. If they cost more than humans per production/hour then why use them? If they cost less than the government/company would be happy to just buy them new shells if it lets them use them instead of expensive humans in dangerous conditions $\endgroup$ – Andrey Nov 9 '17 at 16:44
  • $\begingroup$ Several reasons : first of all, they could be used indefinitely. Humans die of all age and retire or quit the job if they aren't happy. Robots don't. The production costs of these androids can be high, but the use costs low. Humans need salaries and benefits and raises ... Androids need energy and a bit of maintenance from time to time. Think about current industrial machines. They cost more to make than hiring a human, however, companies win back the money on the long run. As I said, androids would be an investment. $\endgroup$ – everyone Nov 9 '17 at 16:59

Not a perfect backup

When you run a backup program on a computer that takes heavy damage, it's not unusual that some file are later missing. Imagine this effect on a powerful AI. The system of backup couldn't be done to save them.

They wouldn't know what happened

Even if you were able to do a "clean" backup without important loss the AI would still have lost the memory of what happend before the last time before saving his memory for a potential backup. It would be really hard to make him the same, because he will miss all of the experience that happened between his backup and the incident.


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