Robots are becoming self aware

[Wolfram23]

I just read the article on SciAm. The short of it is that researchers at Cornell University's Computational Synthesis Laboratory have kind of put "two brains" into a robot. The robot didn't know how it worked but was told to move somewhere. It started testing out it's different servos and such and after a short period of time was able to walk. They then removed a leg, and it was able to re-adapt and walk once more.

Here's the link: http://www.scientificamerican.com/article.cfm?id=automaton-robots-become-self-aware

So, what do you think about this? Where will this stuff lead us?

EDIT: Note that neither me nor the article are talking about complete sentient congnition here! Meerly a basic form of self awareness. Maybe somewhat comparable to an animal. Which is a huuuuge step forward if you've ever worked with/on a robot.

 

[Daystar Clarion]

That isn't really self awareness. At least in the sentience department anyway.

If a fish has a weak fin, it will compensate with it's stronger fin.

 

[Burwood123]

This isn't self awareness, it's as you said, adaptability, eventually we'll have robotic missions to other planets where the robot can operate for very long periods and fix itself. they won't "kill all humans" anytime soon.

 

[Wondermint13]

...I feel stupid. None of that artical made sense to me.

Or was it really just a load of scientific words and robot stuff thrown into an artical to lure a handful of easily influenced nerdy americans into thinking that a robot induced appocolypse is on the way...?

 

[FalloutJack]

I'm sorry, but robots are not becoming self-aware. A robot or computer will be self-aware when it does something like print out "FUCK THIS!" on a turing test or intelligence exam and there's nothing in the programming for it.

 

[Wolfram23]

It's a form of self awareness. No it's not complete cognition but the robot develops an image of itself and how it functions which is a start. Also, they explain how they put the "two brains" into the robot - one is like a beast and just trys to reach it's set goals, it takes care of motor skills. The second one, however, looks at what it's doing and modifies the behaviour.

So no, it's not full self concious but it has some aspects of self awareness.

 

[Gibboniser]

Did it mention anything about being a vanguard to our destruction?

 

[zehydra]

that's not awareness, it's the appearance of awareness.

 

[Marowit]

We have a professor here at our University who studies that too. It's not self-awareness, like someone else said, though.

The way he descibed it at the conference 2 years ago was program-evolution. Essentially the program is given a goal - move from A to B - and given a set of parameters. It is then allowed to crunch numbers essentially, and any movement that doesn't help it move from A-to-B is removed for its list of possibilities. After a weekend of crunching it is able to move. It's essentially a really complicated statistics program (is this action going to increase the likelihood of me achieving my goal?).

It's really interesting work, but you have to control the parameters pretty tightly in order for the program to be able to understand what it's allowed to do and how it must accomplish it.

If I were in computer science it's definitely the field I'd be interested in...just imagine the Cyberwarfare possibilities of having a program that can select against actions that don't help it achieve it's goal --- somewhere in the DoD a General is lying in a pool of drool ---

 

[DefunctTheory]

Its self awareness in a literal sense, not in the actual spirit of what we refer to as 'self aware.'

 

[deleted_by_strongbad]

I could see this as being an initial step that could lead to limited self awareness, which is awesome. I wonder why we are phobic of self-aware robotics/computers. These machines won't be fueled by a fear/aggression based neurology or biochemistry, so the machine's survival algorithms will be largely based upon order of efficiency and gain of benefit. I can't see a reason they would turn on people unless it was a response to mass scale violence against self-aware robots, which if they serve a purpose to us, won't be likely to occur. Even if you pull a matrix and state that maybe humans contain some valuable resource, that resource is most likely more efficiently obtained through other methods, and humans as slave labor would be hilariously inefficient in comparison to the robots building adaptive but non-aware robots for "slave" labor.

Other humans will always provide a much greater threat to the survival of humans than a machine ever will, because the human brain wires survival to the fear and aggression response. The more advanced the AI, the less likely it will be to resemble a human's awareness and that is nothing but a good thing.