Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
I finally had a moment to view this. It is interesting that we are actually going to confront a lot of disruption from computers being better at mundane service tasks than humans, before they even develop true consciousness. The end of the video is the key: computers will be driving our cars, doing our research, and at least most routine aspects of many of the helping / consulting professions (likely, to a degree, including my own). If they can do that without being conscious in the human sense, perhaps the whole question of consciousness in a computer becomes pragmatically moot.
I wish he had given more than 1 slide with no elucidation as background for his closing, concerning how society might need to adapt. My guess is that it is too different, too uncharted a territory, too threatening to entrenched social structures to be adopted before a lot of human suffering is generated. I saw a chart that mentioned things like negative income tax, guaranteed income, separation of labor from compensation. These are several bridges too far for politics as usual. It does violence to the so-called Protestant work ethic ("If a man will not work, he shall not eat").
This trend has been underway in nascent form for a decade or two already. He points out that productivity has gone up while wages have been stagnant or drifting downward. Society has already demonstrably failed to transfer most of this productivity gain to the general populace; it has flowed mostly to corporate level interests. What this TED talk is suggesting is that machine learning will usher in a new era of what SHOULD be ease and leisure and human uplift, but which, with current structures, is likely to destroy the middle class, concentrate wealth, and grind the current down-and-outs into the dust, if we aren't careful. The real threat here is not machine learning, but our dusty old social structures solving the problems of an era that's already bygone.
The end of the video is the key: computers will be driving our cars, doing our research, and at least most routine aspects of many of the helping / consulting professions (likely, to a degree, including my own). If they can do that without being conscious in the human sense, perhaps the whole question of consciousness in a computer becomes pragmatically moot.
I agree that, since "non-conscious machines" will apparently be able to do a large number of these traditionally human-like jobs, there probably will be a pragmatic trend within many AI research and development companies to simply not worry about "consciousness" as such. The dollars will be made by those who develop the most useful machine behaviors and, from many points of view, that will be all that matters. But I can almost guarantee that there will also be some segments of society (philosophers, sociologists, political scientists, human rights advocates...to name a few), who will (I hope) keep pushing questions concerning the nature of conscious experience. People, for example, are going to literally fall in love with machines. Some rather odd folks are already becoming emotionally attached, for example. to "Real Dolls" (life-sized realistic human dolls). Just add some robotic movement capabilities and AI to the Real Dolls and you can be sure that a lot of fairly average folks (not just loony fringe folks) will start to literally love machines as much, or more, than they love people. So people are bound to wonder (and take sides on the issue) "Do the machine love me in return?" "Does the machine have feelings for me?" or is it simply a device that mimics human behavior? As we've seen in this thread, it is incredibly difficult to pinpoint exactly what it means to say that a machine "really has feelings" verses, "merely behaves as if it has feelings," but I, for one, am convinced that the questions surrounding this distinction are not meaningless. Some of the ways in which we talk about these issues could be misguided in various ways, but I still say that the core questions are meaningful and important.
I also believe (speculatively, and thus with less confidence) that qualia are essentially physical processes and thus they have causal significance and thus we can, in principle, develop a theory of consciousness that helps us to identify which types of systems have qualitative experiences and which ones do not (despite the fact that, from a third-person behavioral point of view, the systems might be indistinguishable). I've speculated that the difference between being conscious (qualitatively experiencing life) and mimicking conscious behavior will be found in the nature of the micro-processing that underlies the macro-behaviors. More specifically: Conscious processes will be (at minimum) self-organizing, self-referencing, world-modeling, goal-directed, embodied processes. GOFAI ("Good Old-Fashioned AI") might produce amazingly human-like behaviors in machines, but I predict that GOFAI will hit certain limits that can only be overcome by adopting the types of engineering that I specified as the minimal requirements for conscious experience.
Most of what I've said above is probably not too wildly controversial. Where I really seemed to go off the deep end (from the perspective of several folks here) is when I suggested that a good theory of consciousness will probably requires some modifications at the fundamental level of physics. I still say we need to be able to track qualitative concepts from neuroscience down all the way to the fundamental physical concepts we use to model reality. The problem, of course, is that qualia have a subjective aspect. (Qualia are physical processes and, thus, can be studied objectively via neuroscience and modeled mathematically, but these third-person-perspective models will not capture everything there is to know about qualia because, in addition to the objective aspects of qualia, there are also subjective aspects. Understanding these subjective aspects does not require unabashed mysticism - indeed, the fact of subjectivity follows logically from the self-referential/recursive/embodied nature of qualia (according to my way of characterizing qualia).
The concept of self-reference logically implies the possibility of a unique perspective. Self-reference does not logically guarantee that there is a unique frame of reference/perspective associated with the self-referential processing (thus our ability to imagine "zombies"), but it does imply an epistemological limitation: IF there is a perspective associated with the self-referencing process, then this must be a unique perspective because, if X is a self-referential process, then only X can self-reference X. Other brains can reference X (objective study, e.g., neuroscience), but only X can self-reference itself (embodied first-person qualitative experience, i.e., subjective feeling of being X).
If I'm right, then machines won't be conscious unless they are engineered so that they can be self-referencing bodies capable of instantiating these embodied subjective/qualitative types of processing. Qualia essentially are the embodied acts of this type of system, so a system that does not meet these minimal criteria simply cannot be conscious (according to my theory).
And, finally, in light of the atheism/theism basis of this forum, it may be useful to point out that my theory seems to rule out non-physical or "disembodied" minds. It also seems to rule out the idea that Reality is a sort of "cosmic consciousness" unto itself because it seemingly makes no sense to say that Reality as a whole is a "goal-oriented body" existing in the context of an external world within which embodied actions could make sense. My theory would, however, leave open the possibility of a more or less unitary fundamental "unconscious mind" or "proto-conscious field" or "field of qualitative possibility" that experiences moments of actualized qualitative consciousness via embodied physical processes of the right sort. I think this would be in agreement with numerous mystical traditions, but given the fundamental assumptions, the emergence of conscious beings could nevertheless be mathematically modeled via the principles of self-organization and natural selection (the combination of which could, in principle, explain the emergence of physical bodies of the right sort).
Last edited by Gaylenwoof; 01-06-2015 at 09:29 AM..
wow. It took a video to convince you of what has been said 10's all ready ... amazing.
"proto-conscious field". the term field has definite quantities at every location possible. Some fields can over lap and not effect each other under some conditions. This consciousness you speak of gray is emergent from these field should it exist. Which the data says is more likely "yes" than "no" by the way. But because you don't know of or can't measure a field does not mean it is "disembodied". Think of a year long time lapse picture of a field where people visit but not too often. It looks like they "pop" in and out. They looked quantized. They look disembodied. that is an illusion.
You in essence are coming up with, or assigning, a property that has no carrier of that property. With no "carriers" it cannot be expressed. Well, probably not. I would classify humans as "proto-conscious" because it a dense region in space that is "humanly conscious" but that it is just a sub set of the whole?
Mordant and Gaylenwoof, I take off my hat to you for a reasoned and highly erudite discussion and finally getting to a conclusion that intelligence in a robot (or computer) IS not, in principle, unfeasible, (1) and Gaylenwoof, for seeing its relevance to the atheist debate and keeping it to topic.
(1) and would not be some sort of disembodies floating Mind but would be 'self -referencing', which I take...I've been busy.. .. to relate to being self aware, and the input and response mechanism being in us physical (2) would also be physical in the computer, even if it was made as metals and plastics. In the future we could grow androids in vats. I bags the blonde with the biggies.
(2) I presume that it is agreed that qualia and an explanation of the Hard question is not now regarded as disproving the consciousness could be a physically -engineered product of our brain?
People, for example, are going to literally fall in love with machines. Some rather odd folks are already becoming emotionally attached, for example. to "Real Dolls" (life-sized realistic human dolls).
Yes, there was a guy filmed on TV a year or so ago (I think it was the reality show My Strange Addiction) who had fallen in love with his car and had sex with it regularly. There's a psychological term somewhere for emotional attachment to inanimate objects, I forget what it is. But if we already have that going on, you can take it to the bank that eventually people will be dating / marrying robots. Maybe not even all that eventually. I don't argue that robots will have to be self aware / sentient / conscious in order for this to happen. A car certainly isn't. Many people aren't especially self aware or conscious, come to it.
wow. It took a video to convince you of what has been said 10's all ready ... amazing.
I'm not sure what you mean by this. Does the video say something that contradicts anything I've been saying throughout this thread? I still think that consciousness will not emerge in systems designed on the GOFAI approach, but I've always expected GOFAI to accomplish some amazingly complex behaviors that mimic the behaviors of conscious beings, so what do you think the video convinced me of? If my position has changed, I'd like to know how, so that I can step back and decide whether or not I like the change!
Quote:
You in essence are coming up with, or assigning, a property that has no carrier of that property. With no "carriers" it cannot be expressed. Well, probably not. I would classify humans as "proto-conscious" because it a dense region in space that is "humanly conscious" but that it is just a sub set of the whole?
Funny you should say this because I'm currently reading A Place for Consciousness, by Gregg Rosenberg, and his central thesis is that causation need "carriers" and no such carriers are available given the restraints of current physics (and I agree with him). Notice this is not to say that future physics can't deal with such carriers; it just means that current physical theories don't have the conceptual resources to do it.
Rosenberg advocates a variation of "panexperientialism." Basically, he proposes the idea that micro-elements (e.g., electrons, etc.) can be "experiential" in some fundamental way, without being "conscious," which is close to - but not quite the same as - my proposal for a primordial qualitative chaos. Rosenberg suggests that these experiential micro-systems are the carriers of all physical causation. I've been arguing that fundamentally qualitative, interdependent elements are the carriers of causation, so our positions are extremely similar. But in my view, it makes no sense to apply the term "experiential" to an electron because phenomenal experience only emerges in complex systems (self-organizing, self-referencing, goal-oriented, etc.). Perhaps this could lead you to say that I have properties without carriers, but I don't think this is accurate. The carriers in my system are the primordial qualitative interdependent elements that self-organize (and it's the higher-level self-organizing patterns that take on the experiential "something it is like to be" nature of "experience"). On my view, an electron can no more be "experiential" than an H2O molecule can be "wet." In my proposal, the fundamental qualitative carriers might be hadrons, leptons, etc., but only insofar as these physical entities can be re-conceptualized in fundamentally qualitative (or "proto-experiential") terms (hence my insistence that current physics will need some important tweaks in order to explain the emergence of subjective/qualitative experiences). (BTW: Rosenberg might want to say that his concept of "experiential" is essentially the same as my term "qualitative." If that is the case, then probably the only difference between us is the way in which I characterize the fundamental elements as intrinsically self-organizing and interdependent. So far as I've seen, Rosenberg doesn't do this - although his theory of causation - which is a really cool theory, BTW - might imply self-organization via asymmetrical interconnectivity, in which case Rosenberg and I would be really just saying the same thing.)
At this point I've not read much of the book, so I'm not an expert on what Rosenberg is saying. There is, however, an extensive summary of the book and lots of discussion in the Physic Forum ( https://www.physicsforums.com/thread...ousness.59766/ ) If you start reading this and have trouble finding the subsequent chapter discussions, I can provide links.
Last edited by Gaylenwoof; 01-06-2015 at 11:53 AM..
(2) I presume that it is agreed that qualia and an explanation of the Hard question is not now regarded as disproving the consciousness could be a physically -engineered product of our brain?
I agree with this, once we make a few tweaks at the level of fundamental physics, and, of course, I'm still saying that GOFAI won't be a sufficient approach for engineering consciousness.
I don't argue that robots will have to be self aware / sentient / conscious in order for this to happen. A car certainly isn't.
I agree. People will fall in love with machines whether or not the machines are capable of experiencing feelings of any sort. For some people, it simply won't matter whether the machine feels anything, but we can expect debates to rage. Differing views of ethics, aesthetics, and metaphysics will lead to major disagreements over things like "android marriage", political rights for androids, etc. Sci-fi writers have already covered the bases on this, and I'm pretty sure a lot of these sci-fi issues will become reality within a decade or two.
Even if we manage to develop a good theory of consciousness, people will fight over whether the theory is good, or not. Biological Evolution is about as good of a theory as anything in science, but people (almost all non-scientists) still fight over it. We can count on more of the same if/when a theory of consciousness comes on line.
I'm not sure what you mean by this. Does the video say something that contradicts anything I've been saying throughout this thread? I still think that consciousness will not emerge in systems designed on the GOFAI approach, but I've always expected GOFAI to accomplish some amazingly complex behaviors that mimic the behaviors of conscious beings, so what do you think the video convinced me of? If my position has changed, I'd like to know how, so that I can step back and decide whether or not I like the change!
Funny you should say this because I'm currently reading A Place for Consciousness, by Gregg Rosenberg, and his central thesis is that causation need "carriers" and no such carriers are available given the restraints of current physics (and I agree with him). Notice this is not to say that future physics can't deal with such carriers; it just means that current physical theories don't have the conceptual resources to do it.
I don't need his book really. at T=0 we don't know. making stuff up does not relieve that stress for me. When you talk of a "causation" particle that does not exist given the current state of physics I stop right there. I base what I think is going on what is known, not what is not known. These "causation" particles are all over the place. They are everywhere. That is based on what is known. About 10% of the known universe.
What is known can logically explain "self alignment" at t>0 into working patterns and/or feedback loops. Although keep in mind, we all know that we can be wrong.
As I said so many post ago ... If a machine acts exactly like you do ... it is you.
What words we use to describe it becomes meaningless at that point
I suggest studying proteins very deeply and all they entail. Sorry if you did already. Real sorry! Believe it or not they unlock many idea's that can be taken to the fabric of space itself all the way up to the cosmic web. They are not the end all, but darn do they open up the mind. And they anchor any idea you may have in what is known. Not was is not.
I agree with this, once we make a few tweaks at the level of fundamental physics, and, of course, I'm still saying that GOFAI won't be a sufficient approach for engineering consciousness.
Agreed. Or at least that the problem -solving engineering breif would have to be updated beyond the reel and tape in a Robot's head in the Good old artificial intelligence of Sci Fi. Whatever, whenever and however we find out about the way our cranial stodge produces our mind the bods with hammers and soldering wire will have to find out a way of duplicating it or something that does it a different way, even.
Please register to post and access all features of our very popular forum. It is free and quick. Over $68,000 in prizes has already been given out to active posters on our forum. Additional giveaways are planned.
Detailed information about all U.S. cities, counties, and zip codes on our site: City-data.com.
Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
.. to relate to being self aware, and the input and response mechanism being in us physical (2) would also be physical in the computer, even if it was made as metals and plastics. In the future we could grow androids in vats. I bags the blonde with the biggies.
