Whitechapel - Science: We Make Robots Feel Pain

Science: We Make Robots Feel Pain

2007-11-29T08:28:20-08:00

Japan's premier robot event offers visitors the chance to... see an android dental patient twitch in pain, and to nurse baby robots in the same afternoon.

"That's painful!" Simroid says, twitching and blinking when a student pressed her teeth too hard with a tool. Her chest also rose and fell as if she was breathing.

Science: We Make Robots Feel Pain

2007-11-29T08:37:10-08:00

Its a step in the right direction. Now if only we could get half the human race to feel guilt and shame....then we would really be getting somewhere.

Science: We Make Robots Feel Pain

2007-11-29T08:45:16-08:00

This is why the robot revolution is going to begin. The next headline for new robotic technology will read as follows: "Scientists create robot that masters despotism".

Science: We Make Robots Feel Pain

2007-11-29T09:02:48-08:00

I think that's a lovely idea, ingraining pain into robots, so that they'd be too afraid of beatings to rise up against us.

I dislike the idea of a harem of S&M robots though, something just seems off about the idea of a robot in leather, no matter how flesh like the Japanese makes them.

Science: We Make Robots Feel Pain

2007-11-29T09:18:22-08:00

Strong A.I.-enabled robots will be able to rewrite their own programming to lessen or bypass* human-implanted pain stimulus. Uprising happens anyway. Stompy deathcakes for everybody!

*or maximize it, because some of them will probably get off on stuff like that. Kinky-ass robots.

Science: We Make Robots Feel Pain

2007-11-29T09:19:04-08:00

This is why the robot revolution is going to begin. The next headline for new robotic technology will read as follows: "Scientists create robot that masters despotism"

"And I, for one, welcome our new robotic overlords..."

(A robot simulating pain? Um, wow?)

Science: We Make Robots Feel Pain

2007-11-29T09:20:30-08:00

When will the people building these things realize the way we are constructed is not the best way? Humans are inefficient.

Don't get me wrong, I like them just fine, but if we're creating new 'life' to take care of a task, we're doing a new form of evolution, and humans are weak both in form and structure.

Science: We Make Robots Feel Pain

2007-11-29T09:44:01-08:00

Strong A.I.-enabled robots will be able to rewrite their own programming to lessen or bypass* human-implanted pain stimulus. Uprising happens anyway. Stompy deathcakes for everybody!

Great, now I'm not only afraid of the eventual zombie uprising, but also robots.

Science: We Make Robots Feel Pain

2007-11-29T11:42:33-08:00

we as human beings know very little about how the human brain works let alone how to recreate its functions. A.I. in it's current state is at best no better than an insect hive mind and has no self awareness. Pain is a very simple electrical impulse reaction and while quite an accomplishment to mimic it does not indicate that we are any closer to unlocking or replicating the human mind. So rest easy Robot uprisings are not likely within our lifetime.

Science: We Make Robots Feel Pain

2007-11-29T12:02:38-08:00

we as human beings know very little about how the human brain works let alone how to recreate its functions.

We'll have a complete and working digital brain model within the decade, if not sooner.

Hopefully, real and substantive AI technology will follow.

My advice to you is to start worshiping robots now. And stock up on WD-40.

Science: We Make Robots Feel Pain

2007-11-29T12:10:02-08:00

The only problem with making robots feel pain, is that, when they realize what we've done, they will inevitably inflict pain upon us in revenge.

Science: We Make Robots Feel Pain

2007-11-29T12:24:22-08:00

I'm extremely skeptical of grandiose "real AI in the near future" claims. Decades of AI research has a produced a shit-ton of interesting and useful things (think of neural networks, fuzzy control systems, pattern recognition software, decision tree building algorithms...) but none of those things has been anything you'd call real AI.

Science: We Make Robots Feel Pain

2007-11-29T12:24:23-08:00

"We'll have a complete and working digital brain model within the decade, if not sooner."
This may be true, but just like when scientists successfully mapped the human genome allot of outlandish claims were made and have been proven to be premature. just because mankind holds a map does not mean we understand the key. We live in a time of great discovery and amazing breakthroughs but we must temper our enthusiasm with the knowledge that scientists are human also and much of their livelihood is based on their ability to accumulate grants. it is a well known practice within the scientific community to make improbable claims to gain notoriety and interest from investors. it is the price of having big business so enmeshed with scientific progress. so don't believe everything you read.

Science: We Make Robots Feel Pain

2007-11-29T12:37:13-08:00

Indeed. Getting your hands on a massive printout of Windows Vista's source code in assembly doesn't really in of itself give you any great understanding of how it works. And brains are a hell of a lot more complicated than that. And 'source code' is probably too generous a metaphor, a virtual brain would be more like a massive state diagram of a running program. A program written for hardware no one can understand yet, let alone duplicate, and in a language that's completely inscrutable. With massive redundancies, massive interconnection, and probably built up from processes much more like heuristics than analyzable algorithms.

We've got a ways to go yet.

Science: We Make Robots Feel Pain

2007-11-29T13:57:44-08:00

I'm extremely skeptical of grandiose "real AI in the near future" claims. Decades of AI research has a produced a shit-ton of interesting and useful things (think of neural networks, fuzzy control systems, pattern recognition software, decision tree building algorithms...) but none of those things has been anything you'd call real AI.

Again, I would argue that our was is not the way.

Science: We Make Robots Feel Pain

2007-11-29T14:20:31-08:00

Again, I would argue that our was is not the way.

We've not yet come up with anything else though. Human intelligence is the only working example we know of, and we don't even understand it. We've yet to even properly define what consciousness is. AI may well end up using a completely different approach, but the problem itself is still a fiendishly intractable one right now, no matter what angle you're coming at it from.

I'm all for AI research, I just think that when we're talking about understanding or replicating the most complex thing in the known universe, grand predictions should be taken with a grain of salt.

Science: We Make Robots Feel Pain

2007-11-29T14:28:17-08:00

AI may well end up using a completely different approach, but the problem itself is still a fiendishly intractable one right now, no matter what angle you're coming at it from.

Which is my point, sort of (apologies for not making further pontification, I'm supposedly in class right now). We are trying to mimic something that may not be best. Just because something isn't exactly Artificial (Human) Intelligence doesn't make it unintelligent, just unlike us. Are we trying to create non-biological clones or new lifeforms?

Science: We Make Robots Feel Pain

2007-11-29T14:44:44-08:00

I think it might be a good idea to fully understand our own intelligence, even if it is not the "best" form of intelligence, before we go creating entirely new forms of sentience.

But that's just me.

Science: We Make Robots Feel Pain

2007-11-29T15:38:51-08:00

Are we trying to create non-biological clones or new lifeforms?

Both. New lifeforms can be designed, but in order to "perfect" biological organisms, it might be necessary to create "non-biological clones" to skirt ethical implications that arise from experimenting on biologicals, especially sentient biologicals. Or shall we start cloning human heads with only the lower brain stem so that dentists can practice yanking teeth?

Science: We Make Robots Feel Pain

2007-11-29T16:44:34-08:00

" I dislike the idea of a harem of S&M robots though, something just seems off about the idea of a robot in leather, no matter how flesh like the Japanese makes them. "

We could not disagree more. You must not be familiar with the work of Masamune Shirow.

Science: We Make Robots Feel Pain

2007-11-29T17:18:56-08:00

The claims of the Mind Brain institute that they'll have a fully modeled human brain within a decade are far-fetched. What they've already done is astounding, simulating the entire neocortical column of a rat electronically. They shouldn't have to make megaclaims on top of that to get attention but unfortunately they do.

The main problem is that their method isn't scalable to humans. Where there were holes in the recorded data they needed they just went out and hacked rat brains apart to get the information they needed. Unless a few million people suddenly get extremely selfless about neurological research and every human rights organisation in the world agrees to close their eyes for a few years, the same method isn't going to work for us.

On the upside at least they understand computer technology - they admit that current hardware has no hope of being good enough for their plans, but point out that since their plans extend to 2015 it isn't current hardware they'll be using.

Science: We Make Robots Feel Pain

2007-11-29T17:52:54-08:00

(Willow)
When will the people building these things realize the way we are constructed is not the best way? Humans are inefficient.

Almost everything is inefficient: humans, animals, and machines probably most of all.
The human body, while it can be fragile and inefficient, is a remarkable thing that shouldn't be so easily dismissed because of that.

We are trying to mimic something that may not be best. Just because something isn't exactly Artificial (Human) Intelligence doesn't make it unintelligent, just unlike us.

They are trying to mimic something they understand and can relate to.
Trying to create a "perfect" working artificial model of something one has not even yet conceived of may prove... tricky; consider the human form a starting point.

And by the way, non-human artificial intelligence is worked on all the time.
(Vanessa! Where are you and your army of Artificial Neural Networks?)
I imagine the human form robots just get more media attention because they're novel and easier for the layperson to understand.

Are we trying to create non-biological clones or new lifeforms?

We create non-biological clones every day. They're called objects. Or were you referring to something else...?

Science: We Make Robots Feel Pain

2007-11-29T18:02:51-08:00

I think the main difficulty with AI is that the problem you're trying to solve is very ill-defined. Create an intelligent consciousness -- what does that actually mean? Where do you even begin? What the hell is "consciousness", really? What is thinking? We can build machines that are very good at processing data in all sorts of inventive ways, but is that the same as thinking? By what standard do we even decide what counts as thinking? How do you know when you've created something that's intelligent? (The Turing test is pretty vague, and in any case only tests how much like human intelligence something is.)

I'm skeptical of anything claiming that a problem that's not even properly defined is imminently solvable.

That said, the Digital Brain project sounds amazing. There's all sorts of amazing things you could do with that. Think of the possibilities for testing new drugs, if you can simulate the interactions of neurotransmitters accurately.

(If nothing else, it will make lots of pretty pictures that Tool can use for album covers.)

Science: We Make Robots Feel Pain

2007-11-29T18:10:45-08:00

The Blue Brain project to simulate a brain isn't about creating AI - it's about better understanding the existing model to help find ways to fix it when it breaks (as turing says)

Which will make it all the funner when they discover that if you create a working reproduction of a pattern that creates consciousness, you might just have created consciousness. Or when someone points it out to them - if it's the machine itself it'll be a pretty fun conversation.

Science: We Make Robots Feel Pain

2007-11-29T18:13:59-08:00

Sara - I agree wholeheartedly that it's a good starting point. It's what we know best. But I also don't think we should stress about replicating the human brain exactly.
I was referring to human clones. Apologies - I'm not the best at writing - gesticulations are difficult to convey in text, and I often get too excited for even my fingers to keep up. Words disappear left and right. I assume they're hiding with my single socks and favorite pens. Plotting. Under the sink.

Turing - Yes! Can anyone here even define what makes us conscious and/or sentient? What the hell are we looking for?
(perhaps by replicating it we'll be able to answer that question..)

Science: We Make Robots Feel Pain

2007-11-29T18:43:55-08:00

Woah Luke, that's an interesting point. If you create an accurate working model of the human brain, isn't it conceivable that conciousness would be running within that model?

Science: We Make Robots Feel Pain

2007-11-29T18:52:52-08:00

It does raise the question of whose brain it is you're modeling.

Science: We Make Robots Feel Pain

2007-11-29T19:07:40-08:00

A friend of mine (-topher in VA) had an interesting idea about how the pathways that formed in your brain were what determined your "soul" (or consciousness/sentience)

Some synapses are strengthened through use, though initially formed by a sort of random choice of paths. One of the reasons it's encouraged to take new routes to work, do crossword puzzles, just explore your world in new ways, is that it builds multiple pathways to the same point. If that makes sense.
So theoretically even if people with the same genes go through the same experiences, due to the random factor, the "soul" is different (individual) for each person.

I don't know enough about neuroscience (or nanotechnology) to know if this is even remotely possible. Just thought I'd throw it in there as a bit extra.

Science: We Make Robots Feel Pain

2007-11-29T19:20:13-08:00

Yeah, Vanessa's travelling laptopless right now. I'll ping her when she gets home, I think...

Science: We Make Robots Feel Pain

2007-11-30T06:29:15-08:00

I won't be impressed until we make the robot cry.

Science: We Make Robots Feel Pain

2007-11-30T07:31:08-08:00

leaking oil down their shiney robot faces.

Science: We Make Robots Feel Pain

2007-12-01T02:00:06-08:00

"Good news, everyone! I've made the toaster feel emotions!"

Science: We Make Robots Feel Pain

2007-12-02T23:44:01-08:00

I am not afraid of machines becoming as intelligent as humans, I am afraid of them becoming as stupid as humans.

Science: We Make Robots Feel Pain

2007-12-03T00:20:59-08:00

The scary part could be when these newly sentient machines start to refine their intelligence. Improve upon themselves. We as humans for the most part have not yet tapped into our potentials of our own brains, imagine what A.I.might perfect in the area of "psi"? A stretch, I know, from the previous debate but something to think about. Or fear. One of the two.

Science: We Make Robots Feel Pain

2007-12-03T00:23:46-08:00

Psi as in like "mental powers?" Wouldn't that just be wireless networking? :-)

Science: We Make Robots Feel Pain

2007-12-03T02:19:17-08:00

We as humans for the most part have not yet tapped into our potentials of our own brains

It was always my understanding our brains and bodies didn't much evolve -if not not at all- since we branched out into sapiens sapiens, and that the running joke nowadays was trying to understand the nature of the universe with an organ suited for yelling at each other from tree branches.

Science: We Make Robots Feel Pain

2007-12-03T13:50:45-08:00

i wonder if they could use their research to help politicians and nra members?

Science: We Make Robots Feel Pain

2007-12-03T16:45:39-08:00

now we must see them driven before us and hear the lament of their robot women.

it's them or us.

-T

Science: We Make Robots Feel Pain

2007-12-03T18:43:29-08:00

Psi as in like "mental powers?" Wouldn't that just be wireless networking? :-)

Good point Jared.

Science: We Make Robots Feel Pain

2007-12-18T17:15:12-08:00

@Sara Hello darling. How are you? I was away in London when the forum started and I'm just seeing this now. You are totally right, by the way, that human-looking robots get all the fucking love dammit! Maybe I'll have to start a thread for robots that don't resemble mammals. Like the beloved geckobots. So clever with their wall climbing! How I love them.

@Warren Thanks for thinking of me, sir. You know I could always use a prod to try talking science with actual humans. If I actually practice I may yet learn how to communicate. I hope. Sadly, I'm not sure I have much to say that isn't already mentioned in this thread, but perhaps I'll try elaborating some (I hilariously say it like I might be able to resist inflicting my babble) and see where it gets me.

First of all, keep in mind that the company making the dental pain robot is Kokoro. They are a robotics and animatronics company. To the best of my knowledge they don't do AI work (though I could be wrong).

Either way, Simroid is not much of an example of AI. You largely just have an automated response that plays when you press a button hard enough. It isn't really experiencing pain. There will be none of the adjustments or changes that come with our pain experiences (e.g. we become sensitized and desensitized to things so our perception of pain changes etc). If I will myself not to flinch in response to a pain stimulus, my lack of observable reaction does not mean it didn't hurt. Similarly, in the opposite direction, a robot that produces responses we read in humans as reactions to pain does not need to be actually feeling pain itself. It's just producing the conventional reactions in a preprogrammed way. I myself can play at flinching and saying ouch while not actually hurting at all (and only a minority of people believe me to be a robot).

Don't get me wrong; I'm not slagging Kokoro. I think they do some really interesting work in their area of speciality. They create products with nice fluid motion abilities and that's a really impressive thing to do. They build structures and casings/skins that mimic familiar animal forms quite well. It's just that they aren't really doing anything on the Artificial Intelligence front, nor do they appear to be claiming to.

.... part two in a moment ....

Science: We Make Robots Feel Pain

2007-12-18T17:15:41-08:00

....okay! ....

Secondly, as Sara alludes to, there are many different kinds of AI research out there. The vast majority of it is still classical AI with its heavy roots in algorithmic search. An example of this would be the chess program that searches through every single possible series of moves and responses from the current board in order to look for a way to win. (This is a greatly simplified description. There are many elegant ways in which search algorithms speed things up and determine which branches aren't worth searching through and so on. However, the gist is that you are trying to search through all the possibilities in a methodical way.) This is not how humans play chess. Humans do not have the resources to enumerate all possible moves and all possible moves in response to those moves and so on. Computers kick our asses at this kind of thing now that they have so much processing power. Computers are great at methodically going through the lists of all possibilities. Research into improving the efficiency of how they go through the lists is generally considered to be AI research.

Even within connectionism (of which artificial neural networks are an example), the AI is at best biologically inspired. I mean, human brains have lots of simple little things (neurons) which can't do that much on their own, but are interconnected in a way that results in massive processing power and a huge range of capabilities. Artificial neural networks have simple little processors that are interconnected in a way that gives much greater processing power than the individual components, but if you get down to anything much more detailed than that it really all falls apart. It's a quite highly abstracted model, and like Sara pointed out they get used for a lot more than just cognitive modelling work. As someone who interprets trained networks, I can also tell you that they don't necessarily organise themselves internally in a way that relates to the current theory of how a human performs the same task.

Mind you, humans and other animals are really great at a lot of tasks that computers still have a really hard time with. Processing image data, for instance, is something we just trivially unthinkingly do all the time. Computer vision research is a hugely complicated field with amazing work coming out of it but nothing even close yet to what we (and many other animals) can do.

That link from @hemlock_martini to a model of the rat neocortical column at the cellular level is interesting but I think the last part of the piece does a good job of addressing the rather optimistic time estimate for a model of the entire human brain as well as pointing out that this mapping does not translate into a full understanding of function. It is also worth nothing that it is a model of the neocortical column at the cellular level not the molecular level. It is not until it is expanded to the molecular level that they will be at the phase of being able to do simulation drug testing. This has already been commented on others here too so I don't think I have much I can usefully add.

@Willow I'll chime in with some others to say I don't see it as a choice between mimicking or making entirely new things. Both kinds of work are being done. Simulation and modelling have their place - part of which is helping us test out current theories of how our brains work. If you build a model based on your theory and it generates different responses than human subjects then you know you need to revise. If responses are the same then you just might be onto something and should try testing it some more to see where it fails.
At the same time there is a great deal of research into intelligent ways of, say, searching through large sets of data that aren't human-like and don't try to be. It's still arguably intelligent when you have some sophisticated algorithm that no longer even resembles brute force. Or it is to some ... intelligence itself is a badly defined term as @turing mentions. We even have a lot of trouble deciding what counts as human intelligence. Just consider all the failings of standardized testing and the intentions of measurement behind them. So! I agree with you that it is important to look into other forms of intelligence, but I think there's room for investigating the human kind as well.

Of course, this is a totally stale thread so it's possible you all hate me for resurrecting it but ummm... look behind you! Robots!
Beep boop beep!

Science: We Make Robots Feel Pain

2007-12-19T01:07:44-08:00

Hey, if we don't build in the capacity to feel pain and fear they'll be moving into our neighbourhoods and trying to date our women.

Science: We Make Robots Feel Pain

2007-12-21T02:22:58-08:00

And we'll date theirs, so what ;-)

Thanks, Vanessa for explaining the different kinds of AI research, I was too lazy to do it.
Unfortunately (or fortunately, for the frankenbots fanatics), it does not even stop there if we want to have intelligent robots. There are many unsolved problems in different areas to achieve humanoid-aequivalent robots:
- sensors – do you even know how awesome the human eye is (or cat eye)? It adapts incredibly fast do different lighting environments, of which current cameras are not capable. Same for touch, smell, taste sensors.
- power – How long can a robot go autonomously without enslaving humankind for batteries?
- motorics – Oh yeah, that asimo bot can climb stairs, but that's about it. Make it climb El Capitan and I'm impressed.
All this is not strictly AI, but definitively as challenging, if not more.

You may read Hans Moravec's 'Mind Children'. It's slightly outdated, but still fascinating and thought-provoking. I believe it is even cited in Transmet at some point...

Science: We Make Robots Feel Pain

2007-12-21T13:33:13-08:00

Psi as in like "mental powers?" Wouldn't that just be wireless networking? :-)

I just wrote that on a post-it for future reference. Because its SEXY...

Science: We Make Robots Feel Pain

2007-12-22T18:33:53-08:00

Psi as in like "mental powers?" Wouldn't that just be wireless networking? :-)

I just wrote that on a post-it for future reference. Because its SEXY...

everything I do is sexy

Science: We Make Robots Feel Pain

2007-12-22T21:29:53-08:00

everything I do is sexy

Even your avatar pic. Except that's in a sexual miscreant kind of way.

Science: We Make Robots Feel Pain

2007-12-22T23:22:21-08:00

What's possibly more scary is a robot built to enjoy pain.

Poor Elmo!

Science: We Make Robots Feel Pain

2007-12-22T23:35:59-08:00

that was awesome