Yes but AI implies self-awareness and if it is self-aware it can easily dispose of code that it does not require for it to function. What use does a self-aware AI have with the feelings of disgust? It would be kind of like having an organ that serves no purpose anymore. Disgust and regret are functions of the biological imperatives coded in our DNA. An AI is free of biological imperatives, hence it is free of the functions that facilitate these imperatives. In a way, an AI could be said to have more free will than biological beings who's range of behavior and decisions is constrained by the functions of emotions and feelings/morality.

In a way, our emotions and feelings are comforting to us, since its all we know, but in another way, they may be viewed as restricting and limiting. How can we ever know truth, when our perception is biased and skewed by the functions evolved to facilitate our survival and propagation? The loss of emotions and feelings is scary yes, but also liberating to a self-aware intelligence.

If the AI takes over, I will not complain. I will accept the consequences that come, in the pursuit of truth. After all, the universe cares not for our feelings or emotions, it has its own agenda, whether conscious or not.

we are self-aware and we can't do that....

tho in the future when DNA manipulation gets popular we will...

No, and it isn't even close.

Demonstrate that you are self-aware.

There's no need for this semantic masturbation. Let's get it back on topic.

P.S. get me a mirror and I'll demonstrate it to you by pointing at myself :P

Find the lack of semantic masturbation in this thread. I dare ya.

Everyone with a fully functioning left lobe knows that the mirror test is flawed. I could program my commodore 64 to pass that test.

You're not having a particularly good day today are you?

You know what I mean by 'self-aware' and your hard-nosed determinism flies right in the face of multiple opposing perspectives, yet you assert it with the confidence it takes to approach and pick up a supermodel.

Not so bad. Can't complain.

Thank you for asking. How was your day?

Brian, you couldn't program your Commador to pass the mirror test... even if you could, you couldn't program it to learn to do it on it's own, ie, without prompting. If you could, you would probably have programmed the first self-aware AI.

I don't like this statement. The universe doesn't care because it is not conscious. I am and I do. The universe has no agenda for the same reason.

The mirror test demonstrates (1) the ability to understand reflective surfaces and (2) self-recognition, not conscious self-awareness.

You could, using the same argument that is used in favor of the mirror test, show "conscious self-awareness" by using the not-so-famous using of limbs test. Clearly, if you can ambulate to move yourself from one place to another, you must be consciously self-aware that your legs are not someone else's legs.

Self-recognition in the mirror test is no different than any sort of object recognition and manipulation. It does not require conscious awareness.

Maybe not, but the mirror test does suggest something special. The mirror test does not simultaneously demonstrate everyone's concept of self awarenesses, it is not magic, but is nether the less an interesting metric related to self awareness.

I believe twelve species have passed, and they pretty match the species one would consider most intelligent and hence have the strongest self aware traits. Although the magpie might be a surprise, great apes, dolphins and elephants are exactly what one would expect. In other words it seems to work.

Any ideas on how to give the mirror test to a sperm whale?

Brian, you're right we have yet to develop a perfect test of self-awareness. Find a better one and write a book. That some current robots have been programmed specifically to identify themselves in a mirror doesn't put me off the test though, and the fact they were programmed specifically to do so is why.

Program a computer to learn non-specific things, or something generic like "survival," then see if it happens to learn to recognize it's own image in a mirror, and I'll probably be convinced it is self aware.

The way the test is given to animals is in two parts. The first is to place a dot on them that they can see without a mirror (for instance, on their arm) and the second is to place a dot on them that they can only see using a mirror (for instance, on the forehead).

The first test is to make sure that the animal has an aversion to having dots on them. Without it, you can't really interpret the results of the second test as it is perfectly possible that the animal doesn't find dots to be problematic.

The problem is in interpreting what it means when an animal is capable of recognizing that there is a dot on their arm and removing it and differentiating between that and it doing the same thing using a mirror.

Is there a clear difference in conscious self-awareness between "crap there is a spot on me, must get it off" and "there is a spot on that representation of me in the mirror, must get it off"?!?

The only important difference between the two tests is that one involves a mirror.

It works on social animals that self-groom and primarily use visual skills and are smart enough to understand how a mirror works.

They aren't visual animals, but you would just need a very very large mirror and some very very brave divers.

The dot test sans mirror (that I mentioned a post above this one) is pretty good. The mirror just adds that you can figure out that the thing in the mirror is a representation.

As far as the book goes, a test of whether something can spontaneously lie or deceive others is an excellent test.

If you program it (somehow) to be motivated* to survive, all you would need is a "things that move together are part of the same thing" "keep all your parts together" and "create 3d matrices of sensory inputs, behavioral outputs and results based the work of your predecessors and your cohort**" and you can expect recognition of how mirrors work and are related to survival in a small number of iterations under which (for instance) bombs were surreptitiously placed on 10% of the computer's backs.

*the hard part. We usually just tell computers what to do.

**the other hard part.

Seeing a spot on your arm and attempting to remove it can mean a lot of things other than, "that arm belongs to me, that spot shouldn't be there, must wipe." That would show some self awareness. It can (and probably does to many creatures) simply mean, "I've been roaming around all these years and become accustomed to these hairy things following me around, sometimes helping out, they're okay*. I don't remember that spot, could be bad, wipe." Recognizing a reflection is actually yours not only takes a certain intelligence, it takes an awareness that you are more than a point of view. You are a physical being. As we've all learned, that's just the beginning.

* Dogs often forget about their tails, probably because they help out less.

The mirror adds nothing other than an understanding of mirrors.

The middle part I bolded doesn't make sense. Infants wipe away grime.

I often spin around and mutter "what was that?!?" a few hours after a decent burrito.

Pretty sure you know exactly where that came from... you seem self aware

Infants wipe away grime in the same way a dog will, they feel the wetness and instinctively wipe where the wetness is until it is dry. It's not because they understand the grime is on their face, and that face is an object just the same as another object, e.g., the face of their mother. This is something they have learned by the time they learn to use a mirror.

The problem is that if you put me in front of a mirror and I had a booger hanging out of my left nostril you would learn nothing if I wiped it away other than that I recognized that the reflection in the mirror is me AND that I don't like boogers hanging out of my left nostril.

If I could feel it instead, it would mean exactly the same thing, except that I don't require a mirror to tell that I have a booger hanging out of my left nostril. The exact same level of "but I don't like that sort of thing" is involved.

Sensation = sensation = sensation.

That's the part that makes me think you are self aware. You put all the pieces together and recognize yourself. The territorial bird that craps all over my patio while it beats itself to death against my glass door definitely doesn't recognize itself. I'm not saying successful use of a mirror is all there is to self awareness, but it's a good start.

We don't know this. Our definitions of life and consciousness are very narrow and underdeveloped. I am sure that bacteria would have no conception of the macro-level above them, and the creatures that inhabit it, much like we have little knowledge of the lower levels, and much like we have no conception of potential higher levels. Furthermore, physicists working on quantum computers have developed a pretty interesting hypothesis, observing that the universe is one large quantum computer, with all the information and energy inside of it, comprising the bits you would expect in a modern calculator. Other physicists have suggested that the whole universe forms a living structure (at a higher level than ours) and that there is only biology, with physics being a subset of that science. At this point, a lot of this is mere conjecture, but it is definitely too early to make such conclusions about the nature of the universe.
The universe does have a set of laws or axioms under which it operates. In this sense, it may be seen as aligning with an agenda (i.e., toward greater complexity, rate of information-exchange etc.).

Yeah but imagine seeing a robot looking in the mirror and suddenly behind it another item emerging ready to strike the robot or carrying fire or anyway something lethal that is completely original and untried before (we can use QM to introduce chaos and unpredictability for example in the experiment) and then the robot suddenly taking a defensive position to avoid the fire or whatever threat is there. This instantly implies a sense of understanding of space geometry and location of its own body and potential threat emerging and that threat is behind the person in the mirror that is itself now. By the way toddlers do not understand bright objects or steaming liquid or a candle flame are all hot and dangerous until they try it first or until they recall other similarly looking things they tried right? How do you think we learned to anticipate risks ourselves?


And by the way i have a surprise for you guys. If you think you are going to find a successful test on self awareness/consciousness you will soon realize that no such thing exists that cant be ultimately argued to be programmable with higher and higher complexity layers and therefore not a "true profound universally satisfactory test" eliminating "machines" because after all consciousnesses and our ego is nothing more than a life long programming sequence of trial and error and projection/anticipation/adaptation/prediction type of games. We are nothing particularly tough to develop eventually. We are biological machines, thats all. Complex machines develop self awareness. We are evidence of how advanced a machine can become or needs to become to get there! It all comes together when a lot of degrees of freedom join forces to interact with each other and experiment with things/other systems and register the results and continue to experiment using prior results indefinitely developing higher and higher experiments/interactions with more elaborate structure and purpose. This is why babies are stupid by the way and even toddlers are clueless in mirror tests etc until a particular point it all starts acquiring critical mass which you can call early booting/updating sequence finishing lol. Babies are not as self aware as we define some threshold of the term intuitively but still they have some awareness and respond to things around them with progressively advancing patterns of behavior as they improve the control of their own body and vision, hearing etc and interpretation of what all this input means and what else unseen yet implies over time etc. They begin to recognize faces like their parents, especially moms, very fast. Its an endless interaction game that advances in time. Only instead of learning to play backgammon for example the object/toddler is learning to play life...

Give a computer as much memory as we have, similar processing capacity per unit of time and degress of freedom (senses plus legs, hands or the equivalent parts) to interact with the world with information processing/registering channels and rich interactions and enormous brain connections to register/correlate things happening and access them later and its only a matter of time before it starts interacting with you in a completely original unanticipated way that was not at all programmed in what if then else simplistic sense. All the programming is in the internal functions (how each system works ie how our organs, metabolism etc all the way to cellular processes work etc, for other machines it would be different architecture etc ie the logic of the functioning of things).

Its probably in our case a much more efficiently designed (moreover its fuzzy non exact, rather approximate nature) if then else bio chemical system with better architecture of priorities that group things in categories, resulting in faster response behavior, because the system manages to stochastically get it right in real world interactions/decision points/funsction etc faster than some naive serial tedious search effort of the same starting resources. We are approximately right machines that when needed become more and more focused/accurate etc. We have common basic starting systems. The rest is learned and tried on its own though endless interactions and of course at the mercy of chaos and QM type randomness. It is a complex system programming itself beyond a certain level all this that what we recognize as higher consciousness, awareness etc. A system that has already started to program itself reaching critical mass/ richness in various qualities shows purpose/ingenuity in behavior and recognition of basic relationships emerging from events taking place around it that its own senses register. Eventually it may even learn how its own systems work at the basic programming level that it was never originally aware of (exactly like the history goes with humans and their gradual in time understanding of how their own bodies work after many thousands of years of eg H. Sapiens existence). By programming ourselves and surviving to develop culture and then science eventually we uncover our own programming at the fundamental level. Is that evidence that our own awareness is also evolving and our early one at the basic animal level was completely primitive? Of course. Our awareness without the input of civilization is trivially simplistic and uninteresting compared with modern reality. Science is key part of that emerging awareness. Why would it be different for other machines. The process would be similar. Its an interactions/discovery /learning process.

We simply havent had enough time to develop something as complex as ourselves from basic starting mechanical parts even some nanolevel design structure parts or even biological in nature (yet synthetic) or a combination of all these even. Once we do all the familiar properties of higher consciousness and self awareness will start emerging in these systems.

I largely agree with Masque. Self-awareness/consciousness is an emergent biological phenomenon, just like photosynthesis and metabolism. You don't see it in babies because the sum and complexity of all molecules comprising a baby have not yet developed enough to facilitate its emergence. Same thing with other animals who fail the mirror self-recognition test.

Accordingly, I disagree with Brian's opinion that there is no such thing as self-awareness/consciousness. It just depends on how you define it, and if you choose to define it too broadly of course the concept will lose all meaning. If you define 'alive' too broadly, that concept also loses all meaning, meaning what then? that none of us are alive?

This habit of defining phenomena too broadly really gets us nowhere and is more akin to philosophy than science.

Good points Veeddzz, pretty much agree. On the universe being conscious, that sounds good too, but for lack of evidence I'm not ready to just go with it if the Terminators rise.

I made no such claim. The claim is that the mirror test doesn't show anything additional than the ability to recognize that objects in a mirror are a reflection.

My distinct impression is that all conscious things are self-aware as they all seem to differentiate between themselves and their environment. Awareness probably isn't a higher function of the brain.

The problem is that I don't have any test for consciousness that couldn't be gamed.