What prevents someone from owning a lookalike humanoid robot of someone else? Can't that raise questions about privacy and ethics when say a certain humanoid robot looks and probably behaves like someone near and dear to someone else?

  • $\begingroup$ "Privacy" is an entirely modern construct in human existence. I am not sure if it can be maintained, but because we made it pretty much all this way without it I'm sure we can still go quite a distance if we lose our current definition of it. $\endgroup$ – GrinningX Oct 28 '16 at 19:43
  • 3
    $\begingroup$ While it would be possible to create a robot to look like someone else, getting it to behave like someone else is a different matter - people are by nature unpredictable, so how would you have a robot behaving like a specific person? If this were to be possible somehow, then you could just replace that person with their respective robot, without anyone else knowing, which raises lots of potential problems that are probably more important than privacy $\endgroup$ – Mithrandir24601 Oct 28 '16 at 19:46
  • 1
    $\begingroup$ @GrinningX I dont think that's true at all. We have always had an intuitive notion of privacy, at least as far as recorded history. And the way we define interaction in public spaces , architecture and the way we interact in familial life, takes into account these notions of privacy. $\endgroup$ – user2277550 Oct 28 '16 at 19:53
  • $\begingroup$ How can privacy be maintained now, with smartphones everywhere? It already can not and nobody cares. But your questions seems to be about identity theft, not privacy. Or some specific threat to "privacy" (as if it really existed). Please make it clear which aspect you have in mind. $\endgroup$ – Mołot Oct 28 '16 at 20:17
  • 1
    $\begingroup$ The topic and the question asked do not line up. Humanoid robots and privacy have nothing to do with each other. It does however raise ethical questions which are best asked in philosophy SE. There are also rights to one's image that would be raised, but that's theoretical a law question that would probably best be asked in the Law SE and not what you're asking. Please edit and clarify and/or ask in those other SEs. $\endgroup$ – Durakken Oct 29 '16 at 0:57

Note you have not one but two assumptions here: (a) that the technology exists to build AI humanoid robots, and (b) that the technology exists to build a 3D facsimile of a human being good enough to fool people who know him. Well, maybe also (c) that you can combine these two things.

Clearly (a) without (b) does not create any privacy issues like you're discussing.

(b) without (a) creates most of the same issues you describe. If the art of make-up and/or plastic surgery and/or prosthetics advanced to the point where one human being could be made to look enough like another to fool people, you'd have the same issues that you describe.

This isn't a totally new problem. People have been impersonating other people since the dawn of recorded history. But presumably a robot built to look exactly like the subject person could at least theoretically be better than any attempt to disguise one person to look like another.

All of which is my way of leading up to ... is the premise of the question that we are assuming that this AI robot can successfully impersonate a person so well that even his family and friends can't tell the difference? Or is that part of the question?

If we just take it as the premise that it can be done, then sure, what stops someone from creating such a robot and having him go to your home or office and examine all your most secret things, read your personal papers, etc? At that point I think I'd be worried about a lot of things besides privacy. Like what if this robot steals my stuff? Signs contracts or otherwise makes commitments in my name? Commits crimes that I then get arrested for? Starts arguments with my friends who think he is me and now they hate me for life? Makes love to my wife? Or what if the creator kills me and replaces me with this robot? If the assumption is that the robot can indeed fool everyone, then the only defense is to have laws against it and punishments severe enough to deter people from trying it. But of course there would always be criminals who would ignore the law, and some number would get away with it.

But I find the premise hard to believe.

No matter how good a duplicate these robots are, surely a plot to impersonate or replace someone could be easily foiled by having passwords. If people knew the technology existed, then the fact that someone looks exactly like so-and-so would not be considered definitive. The bank wouldn't hand over money to someone based on recognizing his face, spies wouldn't reveal secrets, your boss wouldn't accept a resignation, your friends would question uncharacteristic behavior, etc. Looking exactly like Fred Smith would prove about as much has having a sticker on your shirt that says "Hello my name is Fred Smith". You'd have to identify yourself in some way that is not easily duplicated. Like a password. Unless the people who make the robots can not only make an exact duplicate of a person's physical appearance but also read his mind to learn his passwords, that would be an excellent defense.

Beyond that, how good are the duplicates? For example, are the robots made of flesh and blood, or are they electronic and mechanical? If the latter, then if someone is suspicious he could demand a blood sample or take an x-ray.

People could ask the robot things that only the real person would know. Like, "What did you say to me when we spoke last Thursday night?" Again, unless the people who make the robots can read minds, there would be all sorts of things that the real person would know that the robot wouldn't.

I think that if someone is suspicious, they could think of any number of ways to tell that this is an impostor robot. If the real person is someone that many people would have a reason to want to impersonate -- if he's a big shot politician or a billionaire or a spy with vital information -- then if people knew this technology existed, presumably they'd make it a routine practice to check. Just like, even without this sort of technology, you don't just walk onto a military base or into a bank vault or even into many business offices. You're expected to have an ID or some sort of proof of your identity that is at least somewhat difficult to forge. The more damage you could do by gaining unauthorized access, the more effort is put into keeping out impostors.

Perhaps you could get away with it if you can avoid making people suspicious. Like if someone walked into my office tomorrow who looked exactly like one of my co-workers, what would it take for me to suspect he was an impostor, even if I knew the technology to make look-alike robots existed? If I tried to talk to him about something we had discussed the previous day, could he bluff his way through the conversation? If he just said that he forgot what we had talked about, I'd probably buy that once, but if he "forgot" everything we've said to each other for the past month, I think that would arouse suspicion. If he said he lost the key to the front door, okay, people lose things. But if he lost the key to the front door, and he lost his ID card, and he forgot his password on the computer network, etc, at some point you'd start to get suspicious.

Even if we assume that AI's can pass for intelligent human beings -- which is a big assumption, we're nowhere near such technology today and one could fairly question if it's possible -- but even if we assume an AI can pass for an intelligent human in general, could it pass for a PARTICULAR person? Could it duplicate all his mannerisms, his style of speech, know everything he knows and not let slip that it knows things he doesn't know, etc? How would the creator of the robot impostor get all this information? They'd have to follow someone around for weeks or months recording everything he said and did. Could they do that with no one noticing?

One could imagine that the technology exists, but is used fairly rarely so that people do not normally consider the possibility. Like, when I go to the bank, they make me show my drivers license before they'll give me money. Does the technology exist to forge a drivers license? Of course. But it's rare enough for people to do it that the bank doesn't worry much about it. If these AI robot impostors existed but it cost a million dollars to build one, no one is going to build one to impersonate me and go to the bank and withdraw the thousand dollars or so I may have there. They might build one to impersonate Bill Gates and go to his bank. The real vulnerability would be people who are worth going to the trouble to build an expensive robot impersonator but who no one really thinks about as being worth the trouble. Someone is a friend of the big-shot general but not really powerful himself, but by impersonating him you can get into the general's office and steal military secrets, that sort of thing.

Whew, long answer. Interesting subject, I guess.


From the Digital Media Law Project:

"In most states, you can be sued for using someone else's name, likeness, or other personal attributes without permission for an exploitative purpose. Usually, people run into trouble in this area when they use someone's name or photograph in a commercial setting, such as in advertising or other promotional activities."


I think a lookalike robot would fall into this area. At the very least, our current laws would provide some precedent to argue this point when such issues come up in the not-too-distant future.

  • 2
    $\begingroup$ A person's appearance can certainly be trademarked. Doesn't the Marilyn Monroe estate still make MILLIONS selling her likeness? Whoops, looks like it expired. Time to get my MM sex doll! :P photosecrets.com/marilyn-monroe-loses-publicity-rights $\endgroup$ – Jason K Oct 28 '16 at 19:59

It all depends on the context of your world. A world I'm building will feature a heavy taboo on AI within human society as the first time AI was created, it spiralled out of control and - in short - caused an apocalypse that left the entire species reeling and barely able to believe that it survived.

The species has recovered since then, and not reverted into barbarism a la mad max; tech levels only dipped slightly and have continued to increase afterwards. It's just that machine intelligence, emotional computing, computer vision and all such topics are considered very seriously before any sort of work is done on them, if at all.

The kind of robots you are proposing, that not only mimic humans in appearance, but also in behaviour to such a degree that someone near and dear to the original human would be fooled would be inconceivable in this setting - unless it was an artefact created before/during the apocalypse, in which case, yes, that's a very big problem (but relatively benign compared to other artefacts from this era).


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