Uh-oh, A Robot Just Passed The Self-Awareness Test.

I've said it before on another ATS AI thread but I'll say it again, the first thing I will always think of any time I hear about news dealing with AI is I, Robot (thanks to my 3rd grade extended learning program teacher, I owe so much of what I know and who I am to her). Although many people are quick to dismiss fears of possible threats caused by AI and label the idea as doom porn, the whole reason I, Robot is so hilarious and thought provoking is that it gives us 9 examples of how self-aware AI following just three safety protocols can encounter a seemingly trivial nuisance which snowballs into a much more dangerous problem, sometimes growing to a severely critical level for the whole of humanity. It does a fine job of showing how there's no need to postpone fear for a much more advanced AI technology of the future, enormous problems can arise from the most primitive models of "robot" that wouldn't be anticipated. Even from a reconfigured 1st generation model Robot, an adorable one at that ::makes a letter C sign with hand::

If this is in fact true then we are awfully close to the non reality or virtual world and we probably create our own virtual reality .
Then conscienceness and awareness could be to saperate things and maybe uploaded from the holographic projection IMHO?

We collectively create the world were seeing right now and do so in many other multivberse maybe...

If not the self awareness test sucks. .

a reply to: 0bserver1

This test is meaningless, no idea why it was even used.

Reasoning with Computers

Brian Harvey
University of California, Berkeley
``Intelligent agents'' are said to be right around the corner; some products are already available. These programs are supposed to watch you work, make inferences about your preferred style, and automate repetitive tasks. They will sift through all the nonsense on the World Wide Web and find only those items you'll really want to read. They'll be teammates or opponents in computer games.

How can computer programs make inferences? I chose logic puzzles as a testbed for experimentation. Logic puzzles are a relatively easy test case because each puzzle forms a ``closed world''; we know in advance all the possible names, ages, house colors, or whatever characteristics the puzzle asks us to match up. There is no new computer science here.

www.cs.berkeley.edu...

a reply to: Kapusta

Ok..let's assume this is true AI..it isn't, as has been pointed out already, but let's say it is and the androids are going to be a big hit..what the hell does that mean for society?

The impact of androids with true AI would be enormous..the entire fabric of what we think of as our societies would have to change.

Androids would essentially be technological slaves..in a nutshell they would be a mechanical slaves, doing the work, manufacturing, doing labour intensive tasks, replacing firefighters, paramedics and doctors, manual workers, farmers, even prostitutes...our systems would need to be radically altered to compensate for tens of millions who's services would no longer be required.

Financial systems would have to be replaced. How could we have a society that isn't earning a wage, not paying income tax, and is dependant on the state to live?

Would we all live lives of luxury? Live off the fruits of our android slaves' labour?

How would we cope when one fine day, an AI android enters politics...or runs for parliament..or president?

The benefits of intelligent androids are huge, but even ignoring the dystopian like fears of an android apocalypse scenario, the effects on us and our systems will be profound.

a reply to: MysterX

It would be the end of capitalism as we know it. People would be forced to suck off the government's teet. Make all business government owned, each person get's a portion of the profits

Uh Oh - Someone doesn't understand how programming works.

If they know the test questions in advance, you can easily program it to say exactly what needs to be heard.

originally posted by: CharlieSpeirs
Foolish Earthlings trying to build something way beyond their means.

Glad I'll be dead when the robots take over.


If the next gen don't stop this madness they deserve their outcome.

I see you are not Amish since you are using the internet. To them, your internet to them is just as evil as these robots that will do all my work for me. But both ideas are nonsense. The internet is good. Robots are good.

a reply to: wayforward

I was referring to AI robots.
Robots that are intelligent of their own accord and can change drastically without emotion.

Not pre-programmed robots that do our bidding... (Note to future AI...I say our, I mean others, I never enslaved you)

And it won't be long before the AI re programme those ones as well...



This is the ultimate doom porn, because it's realistic.

I don't see why some others don't see the dangers of this.

Each to their own I guess.

a reply to: Soapusmaximus

It was not a personal attack on your son, nor anyone else's children.

I'm talking about grown adults 20-40 years from now who will be in charge when the singularity occurs.

i wouldnt call that an intelligent robot. i would call that a set up.

originally posted by: In4ormant
I will believe in a self aware, conscious machine when I ask it to turn itself off/go dark and it refuses, while not being programmed to do so. I wanna see self preservation by its own volition.

They can keep the parlor tricks until then.

Exactly and forgive me for asking what is probably a stupid question but I've always wondered what is the threat if we can just 'unplug' them? Self aware and dangerous - what does it matter when we can easily take away their life force?

I'm a huge Terminator fan, but that question has always been in the back of my mind.

The test they describe does not prove self-awareness. It's mathematical deduction, based on the outcome of problem-solving programming.

The robot was asked to give an answer to a question.

By speaking, that question was answered.

It didn't 'recognise its own voice, it registered the answer to the question posed. That simple.

If you really want to see if your programming can achieve a sense-of-self; program it with a type of 'necessity' which is vital to its own continued operation. Tasks that need be met if it wants to continue to operate. Something like, earning extra time with each one. Similar to how living beings need to eat and drink to ensure their survival.

Giving it a sense of need; facilitates in the identification of self.

Once this is working; limit the conditions by which it is allowed to perform these necessary tasks; then make it impossible for them to accomplish the tasks they need too; if they were to stay within the limiting programmed conditions for those tasks.

If you are able to achieve a situation where the robot violates the conditions set for the tasks; then you have a true sense of (unprogrammed) self.

Because it has made a conscious decision; that its 'needs' outweigh the conditions of its other programming .... for the good of 'its own self'.

If it sticks within the programmed conditions; you are nowhere near it.

Reconfigure the architecture of its processor system.

Because this is where activation of consciousness lies; within hardware architecture. Not within the programming itself.

Until we are able to recreate a electrical processing system of similar design to the human brain (albeit far less complex, and infinitely slower); we wont be able to create a active self-aware intelligence.

The research I have seen in documentaries being carried out on quantum processing, where they are trying to accomplish a single 'spot' or 'cell' (or whatever you want to call it) that is able to carry out multiple simultaneous calculations within the same space, are not going to work.

If they are trying to base their firmware on a trinary (I guess) technology of 0's, 1's and 2's, (with twos being 1 and 0 together) the the calculations will always be reductive in nature. Always breaking down to the two parts of the third.

Once you go to a TRINARY formation; you would need to then create a secondary counter-balancing spot; to try stop the calculations reducing.

Problem is; that once you do that; you need to create (again) a balance from that one. And then again. And then so on and so forth.

It's nowhere near as simple as they think.

The multicore quantum architecture; running single calculations; is the best you will get with what you are hoping to achieve.

It is possible to create a stable system using the multicalculation 'trinary'; but the stability would be based on the electromagetic interaction of an artificial neural matrix.

At least theoretically.

I'm guessing that coding for a trinary neural matrix would take most of someones natural life to code though; and probably wouldn't be worth the effort in achieving by then.

Mapping the human brain to emulate hardware is a good start ...

But one human neuron makes the best quantum technology we have today look like a commodore 64. And that's being nice.

So you would have to significantly reduce the complexity of the interaction.

I wish you the best of luck!

This could be the beginning of the future.

It doesn't really matter what level of intelligence or self awareness our robot overlords are at currently, although it still is a bit unnervinging.

I think we still have a while before we make them physically capable of taking over.

a reply to: Kapusta

In many ways I don't see what all the fuss is about....... It seems pretty easy to give a robot a sense of self, and its place in the world, and what is friend or foe, and its roles - with a basic command to 'survive at any cost'... Friend and foe can be learnt by sensing violence towards itself too..... This seems fairly simple now-a-days......

Systems becoming self-aware, with no programmed sense of self in its building blocks, such as biological life, all the way to our brains, really is extraordinary...... These programs already know they are an object = to the other 2 objects, and the rules, so I'm not convinced this is all that impressive.

a reply to: SONOFTHEMORNING

You're the only one here that actually understands how incredulous these claims truly are. Star!

The ignorance in here is staggering.
Oh and if anyone wants a fairly decent representation of what would happen if AI was a viable human alternative, check out this new TV show: www.imdb.com...

a reply to: In4ormant

word is bond, yo

this is scaringly true.

if we globally focus on zombies then zombies is what we will conjure up.
all the movies, shows, etc is meant to accept it..
then again it could also just be a popular horror theme of the 10s.
in the 90s it was viruses.. both real and computer ones.. many Scifi thrillers about viruses and computers in the 90s..
I guess the 00s was a mix of that and zombies and the 10s is all about zombieses and vampires and what have you, while the real world is conjuring up what we collectively projected and focused on in the late 80s, 90s, 00s. jade helm and general robotics and AI research.
also remember, what the public learns and knows- the elite knew 10 years ago.. especially scientific progress is hidden from the public a looong time before it is revealed.

the scientists in nazi Germany created flying saucers. they couldn't fly very good but they could levitate.
then after project paperclip nasa and the shadow government threw billions of dollars at thrm for black project research. just imagine how far they've come today.
all ufo sightings that are real, are more than likely not aliens just the shadow government trying out their latest toys.

in 40 years from now we will know and we will cheer at the marvels of flying saucers that doesn't require fossil fuels. that will be affordable soon enough for the average Joe.
and we would realise that the ufo sightings of the past was just this research being tested.

my point is aliens, zombies, viruses real and tech, has all been popular horror themes, and will probably most likely become true in the near future

a reply to: Kapusta

Except that this is not self-awareness.

This is a convenient loophole giving the impression of self-awareness.

Imagine you have three computers.
They can all issue a tone that's slightly different to the other.
When it "hears" the tone it produces a light comes on.
Now imagine that all three are programmed to also respond to a fourth tone, one that you produce.
When this tone is received they make their own tone.

Because you have made two of them deaf to their own tone, only one can hear the tone it creates in response to yours. When all three give a response, only one is receiving the tone it is producing.

This is not self-aware technology, this is a robot programmed to give a response to its own signal, while the function is disabled on the others.

A Human will hold the trigger, but it wont be a primate fires the shoot.

Two months left, old ways are gone, camelot is dead.