A robot has passed a self-awareness test

[LazyLion]

A robot in New York has passed the classic King's Wise Men puzzle which serves as a test of the awareness of the self.

The induction puzzle goes as follows:

'The King called the three wisest men in the country to his court to decide who would become his new advisor. He placed a hat on each of their heads, such that each wise man could see all of the other hats, but none of them could see their own. Each hat was either white or blue. The king gave his word to the wise men that at least one of them was wearing a blue hat - in other words, there could be one, two, or three blue hats, but not zero. The king also announced that the contest would be fair to all three men. The wise men were also forbidden to speak to each other. The king declared that whichever man stood up first and announced the color of his own hat would become his new advisor.'

Roboticists at the Ransselaer Polytechnic Institute adapted it for a trio of robots, two of which were told they had been given a "dumbing pill" which prevented them from talking before all three were asked which one was still able to speak.

All three initially couldn't solve the problem and said "I don't know", but when only one of them made the noise, the robot in question heard its own voice and then followed up: "Sorry, I know now!"

The roboticists realise that the completion of this simple test hardly amounts to, as the New Scientist put it, 'scaling the foothills of consciousness', but team leader Selmer Bringsjord said that by passing many tests of this nature robots will build up a repertoire of abilities that will make them become very useful to humans.

"They try to find some interesting philosophical problem, then engineer a robot that can solve that problem," said John Sullins, a philosopher of technology at Sonoma State University. "They’re barking up the right tree."

Robots might still be a long way off achieving 'consciousness' as humans understand it, but their simulation of it can be pretty powerful.

Last year, a super-computer became the first AI to pass the Turing Test, successfully (and worryingly for cybercrime) convincing humans it was a 13-year-old boy.

http://www.independent.co.uk/life-s...s-passed-the-selfawareness-test-10395895.html

 

[LazyLion]

I for one welcome our new robot overlords!

 

[supersunbird]

Roger Roger

 

[Seriously]

Chappie?

It begun!

 

[TEXTILE GUY]

Chappie?

It begun!

For me Chappie invokes memories of chewing gum with useless information on the inside of the wrapper ....... I must be a bally.

Having said that ek glo, nog steeds, Chappie is die Antwoord ...... wat was die vraggie alweer?

 

[Bobbin]

I don't understand why self-awareness is a puzzle or difficult to understand ... maybe I'm naive?

You've got a bunch of passive systems in the brain, and one active system (Prefrontal Cortex). The active system is analytical (Constant survival seeking algorithm). The more advanced the active system is the better it is able to grasp how passive stimuli are affecting or referencing it! "It" being the operative word. If you can point to other things you can point to yourself, it takes analysis.

We're able to hold multiple concepts in our mind, some say up to 7 but depending on subject matter, so WTF is so difficult to understand about analysis? Sure, how it works is a mystery (Somewhat) but we do know it works right?

 

[Bobbin]

The survival algorithm exists thanks to evolution. It HAS to in my view, and every human behaviour proves that it does.

In a debateable order it would function in really basic form as something like this...

1. Stimuli/input. Can be both internal and external, is based on the amount of sensory information and capacity of the brain.
2. Analysis is to assess all present stimuli (Internal and external) and act on priority and/or past experience. (Maslows Heirarchy anyone?)
3. Output is the action.
4. Learning or action reinforcement/or avoidance. (Many ways to achieve this I think) (Pavlov's theory anyone?)

The algorithm itself doesn't even need to be difficult to create a functioning thing!

What makes humans so advanced is the level of sensory information. 5 senses + memories and huge storage capacity is a lot of information to process and assess every second. But in theory the system at its foundation is simple! The goal of AI should not be to replicate a human because in order to do that it would require the same sensory information and be similar in every way in order to reproduce the behaviour. I'd much prefer AI to be problem solving and tuned to specific function. Don't give it what it doesn't need to be useful.

This is also why I don't take AI doomsday prophesying very seriously. The survival algorithm and needs of the machine are explicitly set. If reproduction and world domination are not part of it's logical queues I somehow doubt humanity is doomed. Survival is an explicitly learned property. We need food, resources etc... therefore our survival is set towards those things. A machine would/should! have completely different survival needs.

People need to really re-define Artificial Intelligence by firstly understanding the human mind (And life in general) properly and not pretend it is magic or that a soul exists

Guys, if it survives (Particularly intelligently) it is alive in my definition.

All I'm getting at is this turing test and making it act like a human is quite a load of BS It just feels misdirected to me.

 

[Bobbin]

So when someone who can program well and make that algorithm learn the stock market, please contact me so we can be rich together?

Imagine something that is able to apply those calculations to every variable on the stock market thousands of times a second and all the history available. And learn from it to repeat positive results and avoid negative ones (The essence of survival really, and I could write a lot on this concept and how/why it exists too). Maybe this program already exists hell I'd keep it a secret.

 

[Seriously]

For me Chappie invokes memories of chewing gum with useless information on the inside of the wrapper ....... I must be a bally.

Having said that ek glo, nog steeds, Chappie is die Antwoord ...... wat was die vraggie alweer?

Then were both bally's

Yeah..... Die Antwoord. What a cliche.

 

[Seriously]

I don't understand why self-awareness is a puzzle or difficult to understand ... maybe I'm naive?

You've got a bunch of passive systems in the brain, and one active system (Prefrontal Cortex). The active system is analytical (Constant survival seeking algorithm). The more advanced the active system is the better it is able to grasp how passive stimuli are affecting or referencing it! "It" being the operative word. If you can point to other things you can point to yourself, it takes analysis.

We're able to hold multiple concepts in our mind, some say up to 7 but depending on subject matter, so WTF is so difficult to understand about analysis? Sure, how it works is a mystery (Somewhat) but we do know it works right?

The survival algorithm exists thanks to evolution. It HAS to in my view, and every human behaviour proves that it does.

In a debateable order it would function in really basic form as something like this...

1. Stimuli/input. Can be both internal and external, is based on the amount of sensory information and capacity of the brain.
2. Analysis is to assess all present stimuli (Internal and external) and act on priority and/or past experience. (Maslows Heirarchy anyone?)
3. Output is the action.
4. Learning or action reinforcement/or avoidance. (Many ways to achieve this I think) (Pavlov's theory anyone?)

The algorithm itself doesn't even need to be difficult to create a functioning thing!

What makes humans so advanced is the level of sensory information. 5 senses + memories and huge storage capacity is a lot of information to process and assess every second. But in theory the system at its foundation is simple! The goal of AI should not be to replicate a human because in order to do that it would require the same sensory information and be similar in every way in order to reproduce the behaviour. I'd much prefer AI to be problem solving and tuned to specific function. Don't give it what it doesn't need to be useful.

This is also why I don't take AI doomsday prophesying very seriously. The survival algorithm and needs of the machine are explicitly set. If reproduction and world domination are not part of it's logical queues I somehow doubt humanity is doomed. Survival is an explicitly learned property. We need food, resources etc... therefore our survival is set towards those things. A machine would/should! have completely different survival needs.

People need to really re-define Artificial Intelligence by firstly understanding the human mind (And life in general) properly and not pretend it is magic or that a soul exists

Guys, if it survives (Particularly intelligently) it is alive in my definition.

All I'm getting at is this turing test and making it act like a human is quite a load of BS It just feels misdirected to me.

Youre too clever. Why are you here?............... you should go help the scientists as they are struggling for years already.

 

[Bobbin]

Youre too clever. Why are you here?............... you should go help the scientists as they are struggling for years already.

Exactly!

 

[bokdrol]

well as long as the robot didn't pass wind or pass a kidney stone, I reckon we're still safe....

 

[Ninja'd]

I'd still kick its ass in DOA5

 

[Azg]

"but team leader Selmer Bringsjord said that by passing many tests of this nature robots will build up a repertoire of abilities that will make them become very useful to humans."
Should edit that to "very dangerous" to humans.