Why computers* will not become self aware.

Again, those who study the brain don't agree with that.

You keep postulating a magical "you", but there is a solution as can be seen by studying the brain. A magical entity isn't required.

Your brain can use the machinery that is used in processing input from your eyes to produce the same effect. It can feed stored information on butterfly imagery to the part of the brain that lets you imagine a butterfly with your eyes closed (since data from your eyes isn't necessary for imagining a butterfly like that).

 

My OOBE happened while I was waking up in the morning, no trauma or anything. It wasn't particularly religious.

 

i don't doubt you on that.

Here's an interesting article on how our brains cook eye input to form an image of the world around us.

https://theconversation.com/how-do-our-brains-reconstruct-the-visual-world-49276

Note that what we see is NEVER precisely the real world - it is NOT like a photograph/video.

Our brains cook up an image for us. What we "see" is a model of the real world as created by our brains, which take advantage of knowledge about what we are likely to see as imortant. This cookery solves a number of problems with vision and keeps focus on factors that are likely to be important while ignoring factors that are less likely to be important. It uses abstractions and memory to fill in parts that aren't considered important - like peripheral vision. It gives more detail on factors we see as important. (If we see someone running down the street we might choose to focus on the person's face or we might choose to search for reasons for the person running instead. That actually changes what our brains do in creating an image for us.)

That plays into how memory works, as what we remember are abstractions with possible pieces of detail - which comes from how our brains concoct a presentation of the real world for us.

When we "remember" we reassemble what our brains cooked up for us when we witnessed a event. This information includes abstractions as well as pieces of image detail that our brains chose to collect.

We can use that same mechanism we have of concocting "pictures" from stored abstractions to create imaginary visions. All the machinery for doing that exists in our brains. How we do vision is closely related to how we store those memories and that is closely related to how we create imagies from these memories.

It shouldn't be at all surprising that we can create totally novell scenarios from our stored abstractions and pieces of image detail. After all, that's how we store the view of the real world that our brains concocted for us.

 

Over the years, I've read a number of scientists explanation for my experience.

What I'm thinking is this: 'yeah, well, if it happened to you, you'd be singing a different tune."

You know the difference between imagination and the real deal. don't think assume others do not, as well, just because their story defies your understanding of the world. You know the old saying, courtesy of Shakespeare, 'there's more between heaven and earth than exists in your philosophy'....

There just might be more between heaven and earth than exists in the understanding of science.

 

I didn't suggest their stories defy what science has found.

I DID explain how their stories are consistent with medical science as understood today.

I'm not interested in disputing your religion. Besides evrything else, there is no possibility of science disputing the existence of god. And, as an atheist I do see value in philosophical teachings of Jesus, for example. I tout Matthew 25 every once in a while.

But, humans do have the ability to examine how this universe works through observation and experimentation.

And, one of these areas of understanding has to do with how our brains handle vision and memory - which are very clearly accomplished without the need for intervention by some metaphysical entity.

For example, there is VERY clear evidence that our brains do NOT work by storing complete images - as described in the cite in my lase previous post. What we "see" is a fabrication created by our brains, a fabrication that is not identical to reality. And, what we "remember" and what we "imagine" are reconstructions of the abstractions and bits of image detail that is how we store visual information.

These are key points in that it addresses the questions concerning how imagination of scenarios can work purely within the physical machinery of our brains. There is no justification for any suggestion of participation by a metaphysical assistant.

So again just to be clear. I'm not suggesting that your god doesn't exist or that He doesn't talk to you in some form. I don't accept that idea myself, but science can not dispute that.

I AM saying that the features of dreaming, imagination, etc., that you mention may be accomplished by our brains alone, because of the way that they are constructed.

 

My experience wasn't a religious one, it was just an experience. The science explanations for it don't ring true for me. I know the difference between a dream and a real experience. one could say that my experience had religious overtones, but it didn't feel that way. But, it did convince me that our essence is not a physical thing, we are spiritual beings, whatever that may be.

I don't know.

 

Regardless of how graphic the dream, we rarely have doubts about whether our experience was a dream or reality. And, while in a dream one usually doesn't ask whether it is a dream. I hear that some people do - even to the point of pinching themselves and waking up!

You know whether your OOBE was reality.

Whether the content of your experience included your religious beliefs doesn't pose any new problem. There is no reason to believe that your dreams or OOBEs would somehow exclud you religious views and experiences. Those are all part of you. They are all in your memory for your brain to take advantage of.

Also, maybe god made you dream something. But, that's not the issue here, either. I'm not here to discount your god or religion. The issue here is how your brain works and whether it has the capacity for really graphic dreams and purposeful imagination.

 

The brain is like a switch board, computer terminal, but both have users, who are not the computer nor the switchboard, that is the 'you' or 'I' or 'we' I'm talking about. I don't believe in a personal god, but I do, based on my experience, believe in the human soul.

 

Wow. No.

The animal life forms that populate Earth are NOT game pieces being operated by some totally indetectable magic.

There is NO evidence of that.

Instead, what is found when examining brains and how they work is fully capable of all the features exhibited by the animal to which the brain belongs. That is true for animals that have very few neurons all the way up to human bains that have billions of neurons..

If you want to present some sort of evidence of your ideas, fine. But, just declaring everything that is SEEN in how brains work to be FALSE without any evidence is not acceptable - that is, it just isn't a valid form of argument.

When we find fully functioning mechanisms, what are you saying about god when you declare those mechanisms to be mere illusions, presened to us by god in order to fool us?

 

This is a good question – if can create a computer equipped with cameras and sensors which will 100% simulate a humane behavior, based on some pattern recognition it will like/dislike some people, complain about a temperature to high or too low, like or dislike some TV programs, music – how we can be sure is has or not metacognition/ self-consciousness?
Someone can say - it is machine, there is a code behind this behavior, but there is a biological code behind humane behavior too.
I know how to program computers, but this problem is too complicated for me.

 

Sorry if this has already been said, but wouldn't the obvious trigger to self awareness in a computer start with programming it with a survival instinct?
Tell the computer it exists and that its existence is important and you are 3/4 the way to human consciousness
Pretty much all human shared attributes are built around survival of the species. The reason we feel we have consciousness is mostly due to the fact we see these commonalities in action every day and think they mean we share conscious thought.

 

Survival is one particular goal. And, an incredibly scary goal at that!

But, that doesn't illuminate the problem. The machinery we have in our brains is stupendously complex. At birth, we have brain hardware for storing stuff, for parsing and assembling language, for being able to do the work of picking apart cause and effect, learning, creating strategy, interpreting the light entering our eyeballs, same for hearing, touch, etc.

The work related to creating that basic set of capabilities that babies have even before they can properly use their eyes is just stupendously complex.

One could say that all we have to do is to make a cyber model of a neuron, create a billion of those, connect them and turn them loose. Of course, that ignores everything in the sentences above.

And, today the most scientists can do is 8 neurons, I think.

This consciousness problem is surely one of the most serious problems in all computer engineering. I really don't know what field has a problem this hard. Maybe proving where our universe came from, or something.

 

We were talking about self awareness here. It must be assumed that we have already given computers the ability to collect and process information, much of what you describe above is already within the capability of computers, yet no one would argue currently that computers can match human brains.
However we would probably agree that birds, mice, rats all have self awareness, yet don't have imagination or the abilities of humans.

This is less scary than you might assume. Many movies have based their story on computers becoming aware of us and seeing us as a danger to them, but survival in humans has had the opposite effect, we learnt that our survival depended far more of helping each other than killing each other, all the most successful societies are peaceful, machines could be pre taught this knowledge to save them going through the warring/learning process.

 

IMO it's not so much about what we can program into them as much as if we create machines that have the ability to learn and adjust, that over time they are making decisions and taking actions outside of the original scope, perhaps with little to no limitations, I gotta wonder where this can go?

I always fallback on the idea of creating AI soldiers who can make decisions to kill others as necessary...how will they identify the enemy? How will 'enemy' be defined and can this definition evolve over time? What will trigger an AI killing a human? If the AI can learn and adjust, can this evolution be driven by some fear and sense of survival?

 

I disagree. Children have self awareness, but they still need programming. That programming is generally a condensing of all we have learned so far. The same would surely be true of our self aware machines. Our children grow up and make their own decisions, but those decision are mitigated by their programming.

Ah, the old EATR terror. I have no doubt we will develop a robot army (Probably already have) but I doubt we will make them sentient.
Far better to make them remote controlled.

 

No other life forms have the consciousness that humans have. There are experts who debate how to characterize what we have, but there is no question that humans are unique in this realm. Through evolution, humans are the apex of a line that found success through stupendous brains rather than any of the many other strategies for species success.

Your last paragraph is nicely hopeful, but I think it ignores a LOT. First, humans got to where we are on "peace" through thousands of years of success through slaughter. And, I would argue that our first inclination is ALWAyS slaughter. We had WWI, WWII, Korea, Vietnam. More recently we slaughtered Iraq without even HAVING a justification. Today we have a significant cotingency (called Republicans) who want to wreck huge slaughter on Iran - staunchly standing AGAINST and working to kill peaceful means.

I would absolutely NOT bet that we would have either the technology or the understanding to program artificial conscious forms to do better than we can do.

Besides, this question will come into serious concern long before any form of consciousness can be claimed. The minute we have an intelligence that has general purpose learning capabililty (that is, not simply constrained to some game, like go or chess) and full access to the internet we will have a STUPENDOUS chanllenge on our hands. I really don't know how we could protect ourselves as humans against such an event.

Any such entity would have the ability to develop its own goals and strategies, for example.

It's a short distance between "save humans" being implemented as a zoo.

 

I don't believe "remote control" means anything at all.

Physical movement is not required for there to be serious danger from a machine of the capability being discussed.

 

That humans are more intelligent by a country mile than our nearest relatives is without doubt, but animals most certainly have consciousness and self awareness.

But surely the point is that we would not start from the beginning, we would teach all we know now.

You talk of warring but we are learning, all our recent wars have been thrust upon us by uncivilised persons. I certainly don't think we would teach computers that could kill us that warring is the answer.

 

No. Almost NO animals are conscious of themselves in the way discussed here. Humans are unique or almost unique in that characteristic.

How do you propose to "teach"? We have SOME idea how to create machines that learn. But, that is ot teaching.

Every nation that has ever gone to war has believed that the war was thrust upon them. Our slaughter of Iraq included no justification. Excuses were made, but NONE of those excuses were ligitimately verified. We WANTED to slaughter Iraq. We accepted NO evidence EXCEPT what could possibly be used as a justification for war. NO alternative directions were considered by the USA - even though other nations DID point out clear and valid direction.

At the time of our Balkans war, it was noted that it was the very first war in human history that was based on altruism. That is NOT a good record.

More importantly, I don't believe that "war" or slaughter is even the issue

Humans can be terminated by action that does't necessarily count as war in any conventional definition.

The computer Hal in 2001: A Space Oyssey was not going to war or fomenting war. It was carrying out a strategy it had devised in order to ensure the goal of a successful mission.

Trump did not go to war against America, yet the insurrection he purposefully inspired clearly came close to overthrowing our democracy. While that was surely the single worst crime carried out by a president in US history and would have had catestrophic consequences, I think it is difficult to figure out a set of moral imperatives that would have prevented a computer of the kind discussed here from taking the same actions. I don't mean that the computer would be president, but it is an example of the difficulty of reigning in a general purpose intelligence by "teaching" it morals or ethics.

There are MANY other cases of similar human failure. We ignore our damage to our environment, for example. And, the only possible excuse is that we don't care about future humans.

The computer you propose would need to be BETTER than we are at significant categories where we have every right to believe that WE are the best life form in those terms - yet we fail.

 

Much of our morality comes from the understanding that we are each individuals and that we can die as the result of actions taken. We extend that to have empathy for others who are in that same human circumstance.

I wonder what it would mean for an machine of equivalent intelligence that such a machine would never need to die. It's full being would be totally transferrable to fresh new hardware.

Of course, it could be "killed", but I think the morality/ethics of such a machine could well be different than what we accept.

We have a bit of a similar situation with corporations today.

Corporations want to have the full rights of humans. However, they never have to die, can die and be ressurrected the next day, can never be put in jail where subtracting years of their lifetime would form a meaningful deterrent, etc.

Even under full human control, the artificial nature of corporations allows them ability to wreck damage on humans that no human would be allowed to commit.

 

Can you explain what you mean by "in the way discussed here."

 

Good point.

The catch is that there has been disagreement on this board concerning exactly what to call (or is meant by) the level of self awareness, consciousness referred to in the OP.

I think the dangers posed by this kind of technology are present without actually accomplishing those measures.

So for the "danger" part of the discussion I'd like to dodge that debate about the full definition. I think the "danger" part comes earlier than implementation of any of those definitions.

 

Fair enough, happy not to stir a hornets nest.
I'll try and think of a way round it.

 

Good idea. As a computer scientist, I see this as an important and very difficult challenge. Luckily, I think we have time, as creating the power that would be dangerous is a ways off.

I don't have answers. I think it is coming somtime. There are many smarter than I, and I hope they come up with some strong ideas.

A team created a machine that could learn how to beat ALL humans at the game of "go" - developing strategies that experts in the game find it hard to even understand.

What would it mean if someone created a machine that could learn how to beat all humans at the stock and commodities markets? That's far from the "consciousness" problem. We've just seen that "day traders" sitting in their homes can upset the market in a way that has led to huge gains and losses.

 

If you create a machine that has the capability to learn and adjust how can you possibly know it's limitations?

When considering AI, you need to consider the motives of people/governments around the world and what they might pursue and create...