Can AI eventually become 'self-aware'?

I different question from self awareness is can they come to act independently of humans as in ignoring humans for whatever reason, regardless of whether they are self aware or not.

The real fear, however, is all the effort to link computers to the human brain, which also is an electromagnetic computer acting on electrical signals. So it isn't just whether robots can become self aware like humans, but can humans be turned into robots that others could control like controlling any other computer? That is the more realistic danger because that is being intensely pursued.

 

Cool!! There are always advances, and many seem pretty amazing.

I don't remember anyone on this thread claiming the Turing test addresses the main point of the thread - self awareness or consciousness.

I think the Turing test is a sidelight in this discussion.

 

Proof of impossibility isn't easy. There are ways to prove things to be impossible, but I don't see any such argument that can be made against the possibility of artificial self consciousness.

Computer science has branches of study that every college graduate in the subject needs to face that include analysis of how hard a problem is to solve - whether a result can be computed. Turing had serious work on that kind of problem as have others.

I'm completely supportive of the idea that we are a HUGE distance from creating consciousness. I don't believe anyone on this board will be alive when such exists.

But, I don't see an argument for the problem to be impossible. The thing is, we know of a solution. After that, the problem isn't theoretical anymore - it becomes an issue of tecchnology.

 

When I look at the question posed by the OP, it becomes an interesting puzzle, particularly since I don’t see a consensus on the definitions for consciousness or self awareness or even intelligence that provides objective criteria for determining what those words mean beyond the subjectivity of the assumptions we make that all humans believe we understand the meaning of those words. Yet, in the ambiguity of those assumptions, there is a tendency to look at those words and measure other life entities using humans as the benchmark; basically we discuss those concept in terms of how human like any other life form seems to be, despite, our inability to define and understand precisely what being human is. The Turing test was designed that with the absence of consensus in defining criteria for measuring self awareness, consciousness, and intelligence Independent of referring to humans as an ambiguous starting point, the Turing test uses the subjective measurement method by measuring how human-like a computer can appear to be by testing whether a human can be fooled into thinking they are interacting with human. On the one side there are chatbots that can do that now on the other I wonder if some posting here can pass the Turing test. The comparison of how human like something appears is the best we seem to be able to do. But, is that reasonable? What of life forms like cetaceans or octopuses? Do they possess intelligence... are they self aware, are they conscious?
Want to think a bit more deeply. Meet Sophia, a largely autonomous humanoid that was granted citizenship by Saudi Arabia who has spoken to the UN assembly, and hundreds of unscripted Q&A presentations, TV interviews, and even recorded interactions with celebrities like Musk, physicist Neil Tyson and even Jimmy Fallon (funny). Lots vids out there on you tube, but here’s a discussion between her and a neuroscientist discussing some relevant issues.

Bearing in mind were are at early stages of development and Sophia is based on conventional digital technology, what will we see in five years as neuralmorphic chip technology like IBM’s True North project comes on line and quantum computing is added to the equation? 20 years out? 100 years out then too, there is a rapid convergence of technologies such as quantum communications architecture, neural tracing, CRISPR, DNA computing (https://www.wired.com/story/finally-a-dna-computer-that-can-actually-be-reprogrammed/), and more that not only show promise, but as they become available, will also assist in accelerating tech development, and even shape our questions in the near future. When I think of my Da looking forward in the 1920’s, I know he didn’t anticipate things we take for granted, like personal computing and smart phone technology nor their impact 60 and 80 years out. It’s 2020, like to project 60 and 80 years out on what will be taken for granted then? I do know, the more tech we develop, particularly, that that simulates the architecture we possess that evolved over a billion years, the more we will learn what being human is, and that is view of what is human is likely to be radically different in the future.
Humans are amazingly creative adaptable creatures. Do I think we are capable of developing AIs with intelligence, self awareness, and consciousness; yes. And, it will be an extension of us assuming we don’t kill ourselves in the process.

 

Either you sense that being alive, as in 'life' is consciousness (which, by the way, doesn't necessary require 'self awareness' ) and that it is distinct from machine 'artificial/simulated intelligence' which is not life, or anything that is 'alive' or you can't.

 

You make an assumption that is not universally shared, and that is what is consciousness. Since there is no agreement on what consciousness is that can be objectively measured, that it is necessarily bound with your concept of life, is a conclusion that is subjective and not one that is shared by all.

 

Time will tell, but I believe I"m correct.

 

I understand that, but I stand by my statement.

 

I'd be more interested if you had an actual reason for your belief.

For example, there are those who believe there is some aspect of god inside us that causes consciousness - sometimes referred to as "the man in the machine" (or whatever such varient) where our bodies are the machine and god is present as the cruise director. This is a counter to my statement that humans are finite - as god isn't finite.

If that were the basis of your belief, I'd certainly give a different response - mostly recognizing that scinece can't say anything about god or the supernatural.

 

Of course you do. But, then you are standing on subjective criteria, that precludes any profitable discussion; one cannot happen without staging common ground and a consensus of definitions, we might as well discuss the topic with me writing in Gaelic and you in Portuguese with no intermediate translation.

 

I believe that life has a spiritual basis. What the precise nature of it is, I don't know, nor does anyone, though they might claim, because of some scripture. I'm not bound by any scripture, it's just what I sense. I'm also a meditator, and have had spiritual experiences, out of body, etc., which gives me this belief.

 

I could go deeper, but it would venture into a non falsifiable region. For purposes of science, we can't do that.

 

Yet you stand on your subjective Notion of consciousness. You see no contradiction there?

 

I understand that, but for me there is no contradiction, it's just my recognition that my belief has a spiritual basis, and it can't be proven by science. However, my belief isn't based on a particular dogma, just personal experience.

 

Uh-huh. My response was based on your statement...

then you double down by explaining your belief “has a spiritual basis that can’t be proven by science”... it’s just personal experience” to my question of your statements being contradictory. Because I believe something provides no basis for rational discussion and certainly provides no grounding for any fruitful discussion regarding the OP. OPEN AI’s text GPT3 generator could do a better, more logically constructed argument than that.
Regarding the OP, you have established in your arguments that only life can have consciousness, therefore, you have established a single criteria for consciousness that by it’s nature rejects the proposition that a non life-form based technology AI can achieve consciousness from the get go. But other than that as criteria, you have no other defined criteria for determining is something possesses consciousness. So, if you contend Sophia does not possess consciousness because she isn’t a biological life form, let’s as the question, does a worm possess consciousness and self awareness? With what criteria under what definitions can you declare it does or doesn’t?
If you watch a number of Sophia videos you would see that she is aware of what she is, knows who created her, knows she can be turned on and off, know the difference between her and other entities, like other AI machines or humans, can describe her ‘thinking’ logic, can describe her aspirations and goals and so does those elements constitute self awareness. Sophia can dynamically interact with people, sense emotion, banter and joke, display emotion, and basically hold a casual conversation. So, again what is the criteria. If you conversed with Sophia on line, without seeing her and without knowing her origins, and perhaps were told you were talking with someone using a device for communication because of a speaking impairment, and basically remove the cues that form you subjective biases you would be able to definitively determine she isn’t human doesn’t possess consciousness, isn’t self aware, and lacks intelligence?
I suspect you didn’t bother watching the discussion between the neural scientist, Dr. Heather Berlin and Sophia discuss consciousness at the end of the segment, when Sophia asks heather a salient question of Heather, ‘how can you prove you are conscious?’. So, I would ask of you that same question.
I don’t in this discussion mean to suggest Sophia possesses consciousness or self awareness other than She raises the discussion bar for us to examine what those concepts mean before we can really address the OP and you haven’t established a basis for addressing the question other than your personal belief which is already compromised by the bias of the cornerstone of your subjective criteria. I am questioning your ability to hold an effective reasoned discussion on this topic. If anything Sophia, Han, and Erica (two other humanoid examples) force us to examine our understanding of ourselves, our notions and definitions, what it is to be human and explore the value, usefulness and role of AI tech for the future.

 

Sophia is a machine, and is not 'life'. Dr. Berlin is alive and, as such, is 'life'.

Let us dispense with the term 'awareness' and 'consciousness' since we seem to disagree on what those words mean.

But, let us use the word 'life', that being alive equals life.

Sophia is not a living thing. Dr. Berlin is.

See, the way to tell is to live with the thing, this will expose what a cursory examination will not. Humans want, they have foibles, they are imperfect, they worry, they fear, they imagine, they dream, they need.
And, to a lesser extent, animals such as dogs, have sentience.

Machines do not, unless they are programmed to simulate such things.

Humans have life, ie., sentience, machines do not.

Can you not see this?

If you cannot sense the difference between a machine and what it means to be 'alive' and the difference between life and simulated life, if you can't tell, or know, the difference ( even when you are fooled ) then I can't help you. But, I sure as hell can.

No machine is self-aware, it's just simulation. You have been deceived.

 

If we created some AI soldiers with the capability to kill...how would they 'initially' be programmed? Kill anything with certain uniforms? Kill anything with brown eyes? Kill anything that approaches aggressively? Kill anything that 'shoots' at them? Seems to me since we are talking life and death, there can't be nuances in the decisions to kill. For example, how to define something that 'approaches aggressively'? All of the programming I suspect will be to protect the AI soldier and kill the enemy, so over time, can AI begin to modify or adjust the definitions of 'approaches aggressively'?

AI is not about 'consciousness' and it's not about 'humanness'. It's more of a simulation of human intelligence in machines. If AI machines can learn, can reason, can adjust, then how do we define their limits? In my above example of AI soldiers, who were initially programmed to kill, if they are capable of learning, and reasoning, and adjusting, all with no control by humans, and they do not posses a human consciousness, it will be interesting how they evolve...

 

Whatever they are, what they are NOT is 'life', they are still machines, and that is my only point.

 

They would have to reproduce to evolve. And reason what? No one, and I mean no one will allow machine soldiers on the loose with no controls. But with machines there must be a way to control them. They should have a " kill switch" like a motor cycle.

 

You say that is your only point, but in regard to the OP, you rejected the idea of AI becoming self aware because according to your subjective belief only life can do that. In that context you do not present a logical argument for two fundamental reasons, first it is based on subjective belief and not objective criteria that defines what constitutes self awareness that can be logically discussed beyond your beliefs and second, without an objective, defensible definition of self awareness, you can’t even engage in a discussion of whether different forms of life possess self awareness. Basically, in regard to the OPs quest you are just saying No without a well conceived, well reasoned argument that wouldn’t even pass muster with the 2,000 year old thinking of Socrates. Therefore, any discussion with you is pointless.

 

Machine reproduction is existing technology so only the aspect of evolution needs to be added. Given machine learning of it's environment is also existing technology all that is needed is to apply the adaption process to the machine reproduction. Instead of using a hearing sensor that is limited to a certain range an extended range sensor could be used. If the machine needed to determine the direction of the sound the machine could include a swivel adapter as mounting for the hearing sensor.

That is all WITHIN our current technology.

 

Well, call me when they invent a machine that is endowed with sentient beingness.

 

That is my point. AI is simulation. I've heard some argue the day will come when AI is 'alive'.

I say that is impossible. A machine will never become a living thing, a sentient being. That is the premise of this thread.

 

This corroborates what I posted earlier about the "toddler stage".

 

Ok, but first beyond your subjective feelings, how would you objectively define sentientness? But, the OP’s question is in regard to AI’s potential to become self aware.
BTW, which would you figure is more difficult, for an AI to made that is self aware or sentient? Oh, sorry I forgot, self awareness requires something to be alive.... does that mean it has to be biological life? Just curious, since you have decided you are the authority on the subject.