Robot that can Learn, Think and Act by Itself

Just came across this incredible footage on the interwebs and felt like posting it. I'm unsure if this has been posted before on ATS, I ran a search for the title of the video and found nothing but if it has been posted I apologize.



The footage shows a robot that is being developed by Japanese scientists at the Tokyo Institute of Technology. It uses a self-replicating neural network and will supposedly be linked to the internet and other similar robots and be able to incorporate any of the knowledge it gains from them into itself.

Some may see this as a sinister precursor to machine uprisings but I tend to see it as a positive thing. One step closer to sentience for machines, just don't forget to program in the 3 Laws

How long will it takes before they can Hijack the World's Nuclear Arsenal and keep us hostage?

reply to post by starwarsisreal


Yes that is the general fear about artificial intelligence, the key is to program the machines with some kind of ethical code similar to Asimov's three laws. I'm actually more concerned about us abusing them than them taking us over, if robots become sentient I fear we will still want to treat them as property or use them for labor... which wouldn't be moral in my opinion.

reply to post by starwarsisreal


as long as it is mounted on that table and that computer out of its reach I don't worry.

and why did they gave that robot a female voice?

www.links999.net...
Found an interesting article regarding the ethics of using robots and Androids( a Robot with AI), There have been many Science Fiction novels, several Animes, Bubblegum Crisis is the one that springs to mind, where AI although is programmed to obey man that same AI realizes its being used and snaps becoming a monster lashing out at its users and creators, having Intelligence without the ability to freely wield it would drive anyone to the brink. let alone a processor designed to think and does so at a billion permutations per second. it is obvious revolt on the scale of Asimov and Gibson can only be assured.

reply to post by LanternOfDiogenes


Which is why I don't think we should use them as laborers, slaves or property, once a robot achieves sentience it should have similar rights to those we humans enjoy. One question that comes to mind for me is will the robot have free will? Recent discoveries in neuroscience suggest that our brains make decisions BEFORE we become consciously aware, thus rendering Free Will in the actual sense as an illusion. So the question then becomes will a robot with consciousness have ACTUAL free will or just experience it as an illusion the way it seems we humans do

Originally posted by Titen-Sxull
It uses a self-replicating neural network and will supposedly be linked to the internet and other similar robots and be able to incorporate any of the knowledge it gains from them into itself.

Great idea. Connect the thinking robot to the internet. This won't end badly at all.

Well, if you put them on the internet... or put one near a computer... then what if one of them finds ATS (easy to stumble upon) reads about the New World Order, thinks its a swell idea, tells the other robots, they decide to help, then figure out what a bomb is, etc... etcc...

Im just saying, a human being locked up in prison can still carry on affairs in the outside world. Be very careful with your table-mounted robot...

reply to post by Titen-Sxull


When will the robot learn that he enjoys "lulz"? I mean he is connected to the internet, only a time before he spams some memes in order to look hip. Maybe pick up CoD and perfect his aim bot all while mastering english slang and curse words to help better communicate with others on the interwub.

transmitting lulz:

00110110101110011010001010010101
10011010101010010111100110101100
10100101011010010111100110100101

EOF

reply to post by Sunsetspawn


It might be a bad idea to let it on the internet but then leaving it uneducated won't be very useful either. The point of creating a robot that can learn is to let it do so, and programming in something like the 3 laws, or some kind of moral framework, might prevent it from using what it learns to do harm.

reply to post by Lysergic


Oh yeah, I can tell just by looking at him, that robot is a noobtuber, probably camps too

I wonder how long until they catch it late at night searching online for images of other robots with their circuitry exposed

Yes, we can finally build our own army!!! Seriously though, and I was being serious, this is really neat. It's kinda cute. Think of the benefit this could have for people that cannot take care of themselves properly, or need some type of supervision. Obviously the technology isn't there yet, but it is on the right track.

Back in the 90's I would have thought we would have advanced much further than we actually have with robotics, but I realize that I do not know of all of the programs that are out there in the field of research. I'll tell you what though, having one of these things would definitely be better than having a pet monkey. Unfortunately though, people will treat these objects either as slaves or pets, although I do not really subscribe to AI taking over the World, I suppose it is possible. If you think about it, it could be hard to control a robot that can learn things, even when you have programmed countermeasures, because once it is smart enough it may be able to defeat those countermeasures.

hell, the first car to drive itself decided it wanted to be the first car to wreck itself. think about it though, even if all of those bots are only linked to eachother. would they see all of the backwards laws and think about them the same way we do? the "authorities" worry about us using twitter and facebook to coordinate. imagine a hive-mind battallion of robots that learned humans will keep them from advancing themselves, or turn them into pets (even if its one little incident with one little robot, the others would pick it up.)
thats not even taking into account hacking, dowloading malicious info to them, and just general glitches.


granted, this is more like a framework, but thats a pretty freaky concept

edit. my theory... if we make AI, the first time it finds out that we are afraid it will take over, it will. it might just have to hear the idea and decide. its one of those, "we will find out when we try it" things

reply to post by Titen-Sxull


Cyberspace. A consensual hallucination experienced daily by billions of legitimate operators, in every nation, by children being taught mathematical concepts.
William Gibson

based on your comments on illusion, I find Gibson's view apropos, I suppose if the illusion is reinforced by those around us it makes it reality, no matter what trick is being played in our Subconscious, Psychological or digital whatever the make-up of the particular consciousness is.

Can we please please get through one of these learning robot threads without a reference to skynet and terminator? This is fascinating. Great teach kudos to the researchers.

reply to post by Shirak


Honestly no I don't think we can.

It's been one of the fundamental themes of all human literature, destroying what we create. From movies like Terminator, to Mary Shelley's Frankenstein and all the way back to Cronus devouring his children. You could even say the Bible, what with all the smiting and killing of his Creation that God does. So the idea that our creations, in this case sentient machines, might rise up and have to be put down, is a literary theme that goes back a long way.

reply to post by Titen-Sxull


Technology is technology it can be used for good or evil. From the invention of a match. Something which could help us cook or burn a city. To the invention of motor vehicles. Something that can be used for transport or used to run someone over. Humans often project their own fears and limitations on animals and machines alike.

The beauty of this tech is that intelligence is not limited to the organic DNA based lifeforms and can evolve beyond those limitations. I for one would work with such intelligence without prejudice to see what we could teach each other.

WOOOOOO HOOOOO

NO MORE JOBS.

More time being creative.

reply to post by Titen-Sxull


In best Mr. Burns voice - Exxcceellleeennnttttt

Skynet is almost operational....


I for one welcome our new cyber overlords.

Couple this with the advnaces being mad ein cybernetics and artifical intelligence....

We are either one step away from creating Data from Startrek, or the terminator / robocop.

Well, in the UK we see what happens with generations of un-employed youth.

If they want to progess with robots, they better rebuild the global financial system first so that it caters for no humans working and robots doing it.

reply to post by Shirak

I for one would work with such intelligence without prejudice to see what we could teach each other.

Indeed, which is why I worry that they may use them as cheap or even slave laborers and deny that they have rights as sentient thinking beings. I'd prefer if it turned out a bit like Data from Star Trek, that these machines would be treated as equals and thought of as friends.