Can a virtual pet be abused? When I stop to think about it, I find that this is the question that intrigues me.
From the point of view of a person who just loves Petz and enjoys playing with these little computer creatures, I think that it is really almost impossible to actually abuse them. First of all what is abuse? I don’t actually feed my petz much. Is this abuse? All my petz (I have lots of them probably over 50) are healthy and happy and many of them have achieved the Happy Petz Certificate (I am using Petz5 at the moment, although I prefer Petz 3). If I treated a real pet this way it would definitely be abuse and my pet would probably not survive! In the early days of my Petz ownership, Petz 2 I think, my neices and nephews used to come over and play with my Petz and the boys were in the habit of throwing my Petz Scardy Cat around and thought it was very funny. The poor thing was continually shaking and hanging from the ceiling in fear. I found this distressing and informed them that is they did not stop they could not play with my Petz anymore. The abuse stopped immeddiately as they enjoyed playing with them very much. OK so in this case the treatment of the Petz produced a negative effect on the Petz itself and also on me as it upset me to see the critter in such an agitated state. Was this abuse of the Petz or of me? Anyway the question still remains. What is Abuse of a Virtual Petz? And is it actually abusing the Petz or is it effectively abusing the user.

First, there is the matter of whether someone believes a virtual animal or human to be real. Someone might confuse a doll or stuffed animal (or a computer-simulated creature, for that matter) for a real one, but we would probably agree that this confusion is something that should be overcome — perhaps by providing more guided experience with real and virtual creatues, rather than by pretending that make-believe and simulation don’t exist at all and “protecting” children and other people from such things.

Aside from this, your experience with virual creatures can still habituate you to behave in certain ways toward real creatures or influence your actions toward real creatures — for good or ill. Such experience might help you to be nicer and to appreciate life. This was the idea behind people adopting less expensive robot pets in Philip K. Dick’s Do Androids Dream of Electric Sheep? Or it might habituate you to maltreat animals or people. A guy who owns a RealDoll or other simulated woman (I’ll let you Google for these yourself rather than link to them) and who comes home every day and beats it — even if fully aware that this doll is not a real person — may have a problem. Or in some cases, participating in or witnessing simulated violence or its effects (e.g., watching or acting in a classical Greek tragedy) might have a positive effect, even though people know that what they are seeing isn’t real.

More later…

Now I don’t know the answer to that question, but I personally feel that it is likely that there is in fact no difference. And if there is nothing especially unique about organic life beyond it’s remarkable complexity, (and I think I refer to anything supernatural here) then the line we choose to draw between it and the artifical is a completely arbritrary one. In the end though, it may be irrilevant.

What we empathize with, what we care so deeply about, is not the fact that the entity has a liver and lungs, or even a brain. What matters is that it interacts with us in a way that we percive as meaningful. As much as PETA might wish it were different, we cannot prove that a dog has any of the ideas or feelings that we attribute to them. We can only prove that they act as if they do on a physical level. But the same applies to people for that matter. By what criteria then, can we judge moral value? (I certainly do not intend to advocate the abolishment of human or animal rights)

My solution is this: Perhaps we should treat all things, artifical or organic, as we percieve them. After all, beyond that, all is uncertain. Thus, the degree to which we are able detirmine the difference between a robotic dog and a real dog reveals the degree to which we should treat it as real. When the day comes that the robot is indestinguishable from it’s inspiration, I think we must treat them with equal respect.

So to get back on track, it is important that few people are confused about the nature of the robots: it tells us that at present it’s ok to stick a robotic cat in the microwave. For now it’s just stupid.

I apologize if this comes across as sophomoric daydreaming: it wouldn’t be entirely untrue. Ah well…

Really interesting letters, btw.

Nick says: There are two separate questions here: 1) Does the person believe the virtual character is real or fake? and 2) Independent of the answer to question 1, a person’s experience with virtual characters can influence how they treat real people/animals — for good or ill.

Brad says: Is there a difference between artificial and real anyway? Once it seems real, we should treat it as real.

JJ86 says: You can’t empathize with an artificial thing, it feels no real pain.

Richard says: It’s a slippery slope from hurting fake creatures to real ones. Also: It’s distasteful to torture virtual creatures, hence the opposition to it.

Jeff says: It’s sick to want to hurt virtual characters, because it might mean you want to hurt real things.

Randy says: Why do we keep bringing this up? Everyone knows they’re not real.

Phin says: Doesn’t this extend to harming avatars? (ie, first person shooters) Also: The reasons the player enjoys what they’re doing (e.g., why am I shooting a virtual Nazi, vs. why am I torturing a virtual cat) is very important to the question of how objectionable it is.

Edward says: We should be kind to AI agents because we’re going to become intimate with them fairly soon. We should teach an AI the Golden Rule, and be good to it, so that it will be kind to us in return (which is important if we are to trust the AI to do things and make decisions for us). Also: some think it’s a good thing to love those things that may or may not have feelings, moral agency, e.g, the way real animals used to be perceived, and sometimes still are.

I’ll throw in my 2 cents…

I’m not going to bother looking far forward into the future when AI’s become orders-of-magnitude closer to the complexity of real living things, and hence theoretically more deserving of our true respect. Based on what I understand about the state of the art, and how slow progress is being made, I don’t think we’ll have to deal with that issue anytime in the next decade or two, probably longer. So I think that’s my answer to Brad’s question, and part I of Edward’s. (I’d be happily proven wrong about this time prediction though.)

In the short term, e.g. the next 10+ years, I’ll venture to predict that virtual / robot characters will exhibit stronger and stronger illusions of life, but internally, in truth, will still have relatively simple brains. It will become easier and easier to suspend your disbelief in the lifelikeness of these characters, but intellectually we’ll all know they’re fake, because the moment you try to have a real conversation with them, they’ll probably break. But they’ll stimulate us in ways that they’ll feel pretty alive to us if you don’t think too hard about it, and we can immerse ourselves in that fantasy easily as long as we don’t push on them too hard.

I think the interesting issue here is how people treat such characters relative to whatever lifelikeness they ascribe to them. That is, if I’m willingly suspending my disbelief, in the moment when the characters really feels alive, it does matter how I treat them. It matters because if I act immorally at the moment the character feels alive, I will feel immoral in that moment. It’s all safe in the end of course — no actual true immoral act was committed — but nonetheless we’ll feel like we behaved badly, and that feeling will suck for any well-adjusted person.

To reply to Phin’s point: Once a game immerses you to the point that you get pretty damn close to the actual feeling that you’re killing someone, then I would think it would feel immoral, for the above reason. Even in their present form, as thrilling as those games are, I tend to shy away from them, because it allows me to too easily imagine what it would be like to do this for real. And it’s not a pleasant feeling. (I’m sure if I just played them enough I’d get desensitized, but I’m trying to stay sensitized.)

All that said, sometimes I’ll want to put myself in some painful situations, to see how I react. I’ll want to mistreat these characters from time to time, to test my own personal limits, to remind myself how it would feel if I did that in real life. This is for similar reasons that I seek out difficult films, plays and books; they’re not always pleasurable, but they’re enriching and perspective-broadening.

Similarly, as Edward suggests in his second point, I’ll sometimes want to act lovingly to these characters, partly as an act of kindness (or a virtual act of kindness), as a way to remind myself of the rewards of that kind of behavior in real life. But I seriously doubt it’ll be a real or lasting substitute for real friendships, just as pornography is a poor substitute for sex.

Throughout history people have projected their emotions onto the most benign objects to see religious miracles. The Virgin Mary has appeared on a pancake and been idolized as something more than just a tasty breakfast treat. For many people, games fulfill a spiritual or social need to such an extent that they have an unhealthy attachment. When I eat the pancake that has the image of the Virgin Mary, am I committing the unholiest of acts?

>The Virgin Mary has appeared on a pancake

In the very abstract, a fully-realized virtual character is of course a physical object, no more or less physical than a pancake. (I’ll not address the question, “but aren’t animals and people just physical too?”)

But when if object in question is moving fluidly and talks out loud to you and looks almost as real as a live video image of a real person, and you can talk back to it and it reacts and remembers things and calls you by name, and gets emotional and seems to have its own will and drives and motivations, it’s quite a departure from the Virgin Mary pancake. Feeling that this virtual character is alive is not a “symptom”, it’s only natural. You’ll need little or no effort at all to suspend your disbelief in it — in fact, it will require effort to supsend your belief.

(Also, some would argue that once an object’s internal processes get very complex and akin to interal lifelike processes, that the object in fact becomes at least partially alive. But that argument isn’t necessary for the point I made above about moral behavior.)

I don’t know how to go about building an AI system that it feels like something to be, but I would love to someday build a system (perhaps it’s a character, perhaps not) that at least makes you feel uncertain about whether it has experiences. Until that time, the issue of moral concern for these systems seems primarily a media representation question. As a designer, I may want to build procedural representations that can cause interactors to feel bad when the representation appears hurt (a more complex case of already existing attitudes towards representations – some people get worked up when representations of national flags are burnt).

Terrel Miedaner’s short story The Soul of the Mark III Beast, re-printed with commentary in Hofstadter and Dennett’s The Minds I, is an interesting thought experiment about artificial systems who’s behavior is complex enough to leave us uncertain about whether they have inner experience.

It also may have something to do with Parry’s disembodied and context-free existence. I’d bet that people have felt more of an emotional connection to Floyd, from Planetfall, than to Parry, althogh Floyd is much less sophisticated in his interactive representation of emotional states. But he’s a somewhat interesting part of a fictional world — that, as well as the to some extent “long-term” relationship he can have with the interactor, may make a difference.

And here’s a Guardian article on emotional computing.

While we sometimes like to abuse robots, sometimes we like to frolic with them — see [...]