A Pressing Debate on Pleo
December 6th, 2007 by SBIt’s also a crushing debate, a hitting debate, a debate that is hung by its tail, a debate that squeals and wiggles with irritation and pathetic helplessness.
We are of course, speaking of the recent video of Pleo being abused in a variety of exciting different ways.
Faithful Reader Daneel mentioned the post from Le Boing, and asked “What kind of ethical questions does it raise about the way we treat robots with AI?”
Well Daneel, that is a long, convoluted, and heated discussion fraught with dogma and subjective interpretation of the nature of life and intelligence itself. Meaning, there should be beer on hand before we even think about discussing this in any meaningful way.
On the other hand, you are asking exactly the wrong person, because I must say that torturing Pleos is one of secret guilty pleasures that I do in public all the time, mostly at robot shows, and mostly to show off Pleo’s lifelike response to real-world situations.
Also, Mr. Robotics is the director of partner development for Pleo. He got a Pleo and sent out a bunch to various roboticists and software people. Most of these people had a very strong urge to skin pleo to see all the goings-on and workings underneath.
I guess what I am saying is, people who work with robots all the time tend not to anthropomorphize them, except under very special circumstances. Most robot people are more interested in taking the thing apart and seeing how it goes together, and technology and theory that goes behind how the robot does what it does. They are engineers, not philosophers or science fiction authors.
At this point in time, AI is just that - artificial. There is no real underlying personality or real wants and needs associated with a robot that makes mewling noises, just a sound chip and some lines of code. It’s a representation of life. Eliza doesn’t care if you interrupt her or call her names. Your Aibo won’t care if you leave it off for a long time. Of course in the most telling example, a combat robot doesn’t give a good god damn if its parts are sprayed out all over the arena in a bloody display of crowd-driven blood lust (and the people cheer! you see all the levels going on here?). And yes, combat robots have lots and lots of programming behind them. They are not just R/C cars.
A robot that tugs at the heartstrings and engenders feelings of protectiveness and adoration is really just extremely good coding and product design. But it’s just one step removed from a marionette. With the marionette, you see the puppeteer. With a robot, the puppeteer wrote some code and put it on a chip. You don’t see the programmer like you do the puppeteer, but the robot has no more real feelings than the wooden marionette. If you burn a marionette, no one complains that you’re killing a living thing (sure, you might be destroying a great piece of art, but it’s not a life form.) Robots like Pleo shift the materials from wood and string to silicon and plastic, but beyond that, they’re the same. Which is in no way to say that they’re not valuable as human companions, or that you shouldn’t get them. We at SuicideBots love marionettes. We love puppet shows. We love robots. We just don’t think that when they act hurt, should we as humans respond as though they actually are hurt.
It’s only an illusion of life, a fantasy made real by the puppeteer and his audience. Two steps removed from an actor playing Hamlet on stage (he’s not really dead at the end.) The software engineers behind robots trick you into empathy just as would Sir Laurence Olivier on stage or the way David Copperfield tricks you into believing that the Statue of Liberty disappeared (it didn’t.)
As an adult, you see a teddy bear for what it is - a cute bundle of cotton and paint. Lifeless. Cute, but without soul or feelings. But try telling that to a three year old. To them, it’s just as alive as a Pleo is to you. And if I smack the Teddy Bear, little Suzy will cry - but Teddy won’t. Her feelings were hurt, but Teddy’s weren’t (because Teddy has none). So when Pleo is tortured, some feelings may get hurt, but they aren’t Pleo’s. Again, everyone should do buy one. They’re incredibly cool (again, disclosure - Mr. Robotics consults for them.) But “cool” does not equal “sentient.”
Of course all of this is written with the caveat that if Johnny 5 comes up to us tomorrow and says “No disassemble!” we’re not going to laugh and fire a shotgun into his face. That would be rude. Our technology, however, is not there yet.
As I said tis debate is fraught with nuance and the human condition, so I am sure this will piss someone off somewhere.
Comments?














December 6th, 2007 at 11:57 pm
Yes, yes, yes… that’s all true. But, you’re missing the point. The real pressing question about Pleo is “Will it blend?”
December 7th, 2007 at 5:10 am
[...] A Pressing Debate on PleoBy SBAI Evil Mad Scientist expressive robots half baked ideas RoboToys robots spazout teh ky00t therapy uncanny valleyAI, Evil Mad Scientist, expressive robots, half baked ideas, RoboToys, robots, spazout, teh ky00t, therapy, uncanny valley.Suicide Bots - Hot Bot On Bot Action! - http://www.suicidebots.com [...]
December 7th, 2007 at 9:29 am
No Lem, I think the more pressing question is, if Chuck Norris stared too hard at a Pleo abuser, would the abuser have severe internal bleeding or would they just explode?
December 7th, 2007 at 2:21 pm
Putting aside for the moment the awsomeness that is Mr. Norris, this reminds me of a type of Turing test I believe is called the Chinese Box problem. Actually I guess it’s the inverse of the chinese box problem, because instead of proving the equivalency of the contents of the chinese box (ie, whether there is a person inside the box who “understands” chinese and responds, or a person who has no understanding of chinese but manipulates symbols which happen to be chinese and replies with equivalent answers as the chinese speaker), we see an equivalent emotional reaction to an object that may or may not be “sentient”. If my puppy doesn’t really feel pain, but “acts” like it feels pain, is it cruel to kick it?
Maybe the answer is that one “should” be nice to anything that’s on this side of the uncanny valley, but then raises some unfortunate questions about some humans I know…
Anyways, I took a class from Douglas Hofstadter when I was a wee undergraduate, and had to buy “The Minds Eye”, a collection of essays edited by him and Daniel Dennett. Good reading for humans and and other sentient beings who ponder such questions as “does my teddy bear really love me?” and “how the heck DO we get around the Godel Incompleteness Theorem?”
Love the blog!
bf
December 9th, 2007 at 1:43 pm
Your argument starts with the assumption that there is a difference between acting hurt and being hurt. I suppose if I punch you in the face, you’d prefer me to believe that you are hurt, rather than simply acting hurt, but how would I know the difference?
Moreover, doesn’t it say anything about me that I can go around punching people in the face, irrespective of the fact that those people may or may not be able to feel pain? In the same way, what does it say about people who can inflict wanton damage on a “representation” of life (your terms) for their own enjoyment, simply because it doesn’t “feel” pain. Your fictional engineers resort to vivisection as their initial response — and so would biologists when they see rats do something biologically amazing — that’s a statement about the engineers and biologists.
Simply, what is the difference between “feeling” pain and “acting” hurt, and how can you prove to me that you feel it?
Jesse.
December 10th, 2007 at 11:37 pm
Funny that Mike K. had a similar post on a similar subject on the very same day
http://www.orangecone.com/archives/2007/12/roomba_intimacy.html
Frankly, I think Dennett nailed the whole problem with his “intentional stance” business: the human mind is *hard-wired* to treat complicated things as sentient even if they are obviously not, i.e. “this coke machine must hate me…” When things are designed to be cute and cuddly that blurs the line even further.