> > Let's play thought experiment. If, instead of the toy-level robot pets we have today, we sometime down the road create a highly sophisticated robot pet that looks and acts just like a real dog, but with no messes, no health issues, no need to feed,and no possibility of running away or attacking somebody or what-have-you, would that still not be as good as a real dog? If not, why not?
>
> No matter how realistic you make it, some people will still feel a difference because it is not alive. Being living things ourselves, we tend to empathize with other things that are also living, especially if they behave in a way we can attribute to personality and character. If a tractor is destroyed, someone might feel a sense of loss, especially if he or she has worked with that particular tractor for a very long time. However, the feeling generally is not crushing, and if provided with another, equally capable tractor, the owner would most likely not care one way or the other. However, if a pet dies, the owner can experience a deep sense of loss, and acquiring another pet usually cannot ever fully replace the old one. Such is the difference between what a dog means to a person, and what a robot means to a person. If your robotic dog breaks or malfunctions, you can always get another one just like it. If your dog dies, it's usually irreplaceable.
>
> knivetsil

One could argue, though, that in the "sometime down the road" time frame Darien's thinking of, robot dogs will be as distinguishable from one another and as *impressionable* in terms of behavior and personality as biological dogs. So if you lost your robot dog of fifteen years, simply buying a new model of the same type wouldn't cut it. You wouldn't have that particular dog which will respond to you in the particular quirky and special ways that made it a great pet -- instead you'll just have a comparable set of hardware and base behavior (assuming no technological progress over those fifteen years, which is ludicrous). In other words, the same sort of situation you'd have trying to replace a lost biological dog with a new one of the same breed.

Of course, by then there'll probably be ways of "backing up" the robot dog's knowledge base and personality, but that doesn't lessen the maximum potential loss -- it just makes it less likely that the loss would be permanent.

No, I don't think the biggest problem is that people aren't willing to believe that robotic pets can be (or at least eventually will be) capable of being unique in some deserving way. I think it's that it feels de-humanizing or degrading in some way to feel an attachment of this kind to something unnatural, living or otherwise. [I strongly suspect this is analogous to a general feeling of contempt for the prospect of shifting one's activities to virtual worlds, when the technology becomes available to make it feasible.]

Issues like this over pets will be just the first wave. Then we'll have the whole mess with the ethics surrounding artificially enhanced humans and strong AI (e.g. "Yeah, it talks and reacts like a human, and it seems to think in the ways that humans do, and it seems to go through the same sort of things humans do following certain emotional patterns, and it reacts to stimuli we'd call painful in the same way, but how can I really *care* about it if I can just throw it out and get a new one?")