Japanese Scientists Unveil Thinking, Learning Robot

[SomeBritishDude]

"Ok, now let's tell the robot to pour water into the glass"

"I'll pick up the glass...I'll pick up the bottle...I'll make the puny human pay for patronizing me with menial tasks..."

Seriously, we're fucked.

 

[008Zulu_v1legacy]

For everyone here, what is this fear of robots and AI you all seem to posses? Where is the real world data to support these outlandish claims that robots will rise up and kill/enslave us all?

I'm liking the idea of a robot that can learn.

 

[Viral_Lola]

By your command!

 

[Lenin211]

"Robot, could you get me some water?"


"I'm sorry Dave, I'm afraid I can't do that."


It will be awesome to have intelligent robots in the house!

 

[Tharwen]

Well done, Japan! You have some clever people.

I think a lot of what it was doing then was manually programmed though... for example, it already knew each of the sub-tasks (e.g. pouring water, picking up the cup), as shown by the recording describing what it was doing.

 

[thelonewolf266]

Never mind the possibility of them enslaving our race or murdering all of us I wouldn't connect it to the internet because there are a lot if very dirty very wrong things it could learn that I do not want it attempting to do any where near me.

 

[MegaManOfNumbers]

Wait, "It also uses the internet to learn"

Shit, Rule 34 with robots are about to become a reality.

 

[Vigilantis]

It connects to the internet to learn? Hope it doesnt visit the Terminator 2: Judgement Day IMDB page.

Dont expect humans to survive another 200 years if we keep this up

 

[A Satanic Panda]

When does something because sentient enough for us to consider it equal? We will essentially introduce a non-human creature (artificial or not) in a human society with the ability to understand it. But it won't have the history and evolution of it's own culture to take from and contextualise. Try explaining why it can't have a girlfriend when it reads so much about them. Then try explaining why it's going to become obsolete in 3 years time, yet remain functional and aware for many years after that.

Making something more human when it is not human isn't a good idea. If this learning is limited to practical tasks and environmental interaction, then fine. But if artificial intelligence develops it's own true consciousness, it could spell a whole host of trouble.

Also... hivemind........ why?

I'd say emotion would be the deciding factor. And I don't see that magically happening like it does in the movies. A robot could replicate emotions if it learned what situation they'd apply and such, but it couldn't actually feel them.

As for the hivemind, I think that's the coolest part. This is definitely a huge step, but we've had somewhat adaptive AI being worked on for a while, none resulting in something quite like this but still. The hivemind thing means they can learn much faster. It would be horribly inefficient for every robot to have to learn the same things.

So yes, I am incredibly happy about this. It's made my night (well, morning really).

Yea, emotion would be an issue, it's not like it could replicate the exact chemical reactions required. It could as you say, initiate emotion. But really, is emotion the deciding factor. Emotion is not indicative of sentience. Intelligence and freedom of thought is more of a factor for me. They're going to eventually become smart enough to notice the societal gap between how they and we are treated. Just because they can't cry about the divide wouldn't mean they didn't understand it. There is also no saying they won't develop some digital equivalent of human emotion. Not chemical but an unconscious number crunch that involuntarily simulates emotive response.

But what stops SOINN from filling the same shoes as neurons? After all what it all boils down to is processing information. That's what our brains do, but our brains are very nonlinear about it. But whats to say that soon robots won't feel emotion like we do? just with 1s and 0s and not with Serotonin?

 

[danirax]

so anyone else here smell technical failure, robots taking over mankind, mankind sends some one into the past to fix all this....or is it just me?

 

[Saelune]

If not everyone, they will atleast take out bartenders based on this.

 

[Frozen Donkey Wheel2]

So...it's exactly like the Geth?

Also, did anyone else think that music in the background was somewhat Portal-esque?

 

[Mirror Cage]

Finally my wise and future-thinking investment in building an EMP-armed bunker below my apartment will pay off! I'll just go triple-check my inventory of bottled water and emergency rations.

So as one bunker owner to another, do you prefer to eat canned chili or SPAM for the rest of your natural life?

 

[Dominic Burchnall]

Well, who'd have thunk it? We're gonna beat the Quarians to it. Hope everyone's prepared to live in enviro-suits for the next 200 years or so....

By the way, I like iRobot, one of my favourite films, but just to continue the Mass Effect 2 trend;

"Mistress Hala'Dama. Unit has an inquiry."

"What is it 431?"

"Do these units have a soul?"

Food for thought.

 

[Ruwrak]

Makes me wonder if they named it Glados...

 

[A Satanic Panda]

When does something because sentient enough for us to consider it equal? We will essentially introduce a non-human creature (artificial or not) in a human society with the ability to understand it. But it won't have the history and evolution of it's own culture to take from and contextualise. Try explaining why it can't have a girlfriend when it reads so much about them. Then try explaining why it's going to become obsolete in 3 years time, yet remain functional and aware for many years after that.

Making something more human when it is not human isn't a good idea. If this learning is limited to practical tasks and environmental interaction, then fine. But if artificial intelligence develops it's own true consciousness, it could spell a whole host of trouble.

Also... hivemind........ why?

I'd say emotion would be the deciding factor. And I don't see that magically happening like it does in the movies. A robot could replicate emotions if it learned what situation they'd apply and such, but it couldn't actually feel them.

As for the hivemind, I think that's the coolest part. This is definitely a huge step, but we've had somewhat adaptive AI being worked on for a while, none resulting in something quite like this but still. The hivemind thing means they can learn much faster. It would be horribly inefficient for every robot to have to learn the same things.

So yes, I am incredibly happy about this. It's made my night (well, morning really).

Yea, emotion would be an issue, it's not like it could replicate the exact chemical reactions required. It could as you say, initiate emotion. But really, is emotion the deciding factor. Emotion is not indicative of sentience. Intelligence and freedom of thought is more of a factor for me. They're going to eventually become smart enough to notice the societal gap between how they and we are treated. Just because they can't cry about the divide wouldn't mean they didn't understand it. There is also no saying they won't develop some digital equivalent of human emotion. Not chemical but an unconscious number crunch that involuntarily simulates emotive response.

But what stops SOINN from filling the same shoes as neurons? After all what it all boils down to is processing information. That's what our brains do, but our brains are very nonlinear about it. But whats to say that soon robots won't feel emotion like we do? just with 1s and 0s and not with Serotonin?

Did you even read all of my post?

Particularly the part I bolded?

rereads Gosh I feel stupid now

Disregard that post

 

[Rin Little]

Robots like that scare the hell out of me...

 

[Garchomp445]

Yes! We might have androids too, soon! That would be so cool! I could have a robot pen pal! Or a robot-dog or a robot-house or... Gah the possiblities are endless! You could have your entire dining room talking to itself!

 

[SeriousIssues]

It also might become addicted to My Little Pony.

I can understand the notions of Skynet and Terminators, but a robot liking My Little Pony is just ridiculous.
...
...
I mean, scientists can fuck up that badly, right?

 

[algalon]

Internet search: Upgrades. Newegg will make a small fortune once this robot makes the logic leap that in order to complete some tasks it will need to learn more complex actions.