Could Robots and Humans Live Side by Side?

My post here is mostly an expansion of the "Transcendent Man" topic that already exists in this section of the forum, and also is a general response to exactly how AI will advance in the future and the collapse of Moore's Law (as this was never answered in the other topic). I have never watched the Transcendent Man documentary, so bear with me if I repeat anything expressed there. I do believe the documentary focuses more on changing humans, and does not explain exactly how technology will advance, only that it will. This is my explanation.

http://news.yahoo.com/video/tech-15749651/could-robots-humans-live-side-by-side-27748600.html

We already exist side by side with robots...but can we be 'friends' and live with them instead of them just existing together? Should we even try to give them a type of 'life'? And how will this come to fruition?

This could mean so much for the future of the worker...one day perhaps even making human labor obsolete.

It has become so much more than just augmenting ourselves to be robotic. In fact, self augmentation seems like child's play compared to the possibility of robots becoming self aware, being that we are naturally conscious beings while they are without this skill.

Pictured above is Moore's Law. Moore's Law illustrates the seemingly perennial trend indicating the growing number of transistors able to be placed onto a microchip (integrated circuit).

The fact that we are rapidly advancing in this field of science will without a doubt lead to a multitude of new technology in the coming years. With the dawn of a new year, we grow a step closer to finally being able to make recursive self improvement (Seed AI) a reality, though seemingly far fetched.

Those who doubt the continuing development of AI are wrong. Those who think it will continue to grow the way it is growing now are also wrong. AI will advance exponentially, and this will happen without the use of the modern silicon microchip.

Moore's law itself is soon to be obsolete. Silicon (building block of integrated circuits) is unstable at the molecular level. If the transistors were to get small enough, and powerful enough, they would ultimately melt from the amount of heat generated. Electrons would leak out and become lost (see Heisenberg's uncertainty principle). Replacements for silicon are the future of the PC, the robot, of everything in terms of technological advance. Quantum computers? Optical computers? Molecular? What's next? And can THESE provide the means for anything resembling singularity?

While some might argue robots are already smarter than humans...this is not true. Intelligence is roughly defined as the ability to understand new concepts and think abstractly, neither of which robots can do at present time. Robots can see much better than we can, but don't know what they are seeing. Hear, but don't know what they are hearing. But can these future possibilities provide the means?

Using atoms (quantum computing) rather than silicon could ultimately revolutionize the field of AI.

It is quite the wild card we have here; do not be so quick to cast aside the possibility of something akin to singularity.

Technological advance (evolutionary systems, for that matter) has proceeded at an exponential rate, and what is more, the rate of change is itself increasing exponentially. Moore's law describes a paradigm in fundamental technical approaches, but Moore's law applies only to semiconductor circuits, and it will eventually reach the climax of its S-Curve (specifically, for existing computer technology). At this point, a new paradigm shift will extend the exponential growth rate of computational devices (i.e., independent of integrated circuits); in other words, Moore's law will continue to obtain in a very general sense, although it will accelerate noticeably. The future "sixth paradigm" will likely be propelled by three-dimensional molecular computing (also known as DNA computing).

As an aside, I do not necessarily agree with all of Kurzweil's extrapolations, but I do believe that it is vital that we reflect upon his theory and its profound implications for humanity.

Altair wrote:While some might argue robots are already smarter than humans...this is not true. Intelligence is roughly defined as the ability to understand new concepts and think abstractly, neither of which robots can do at present time. Robots can see much better than we can, but don't know what they are seeing. Hear, but don't know what they are hearing. But can these future possibilities provide the means?

While I don't doubt that incredible technological advances are on the horizon, the notion that they are inevitable is fallacious. There are significant economic, cultural, and ecological barriers which will have to be overcome before we come anywhere near human singularity. Regarding artificial intelligence in particular, we don't even possess a full understanding of how the human mind functions. Artificial intelligence researchers, for example, have attempted structuring robotic minds around the popular massive modularity theory of human cognition (which evolutionary psychologists and certain cognitive scientists espouse), and the results have been found lacking, to put it mildly:

"...AI has failed to provide a serious account of a routine cognitive capacity, like, for example, getting across a busy street without being run over; or figuring out where to have lunch; or following an argument (to say nothing of inventing one), or playing chess; or ...etc. Simulations of cognition are, in effect, experimental tests of the architectural assumptions of the theories they embody. So far, when the architectural assumptions of Classical cog. sci. have been applied to the problems of commonsense cognition, the experiments haven't come out well; adjectives like 'appalling' spring to mind."
Fodor, J. (2005), "Reply to Steven Pinker 'So How Does The Mind Work?'" Mind & Language, 20: p. 28.

Aside from the possibility of achieving artificial intelligence, I'm undecided as to whether I even find the concept desirable. I do support the further automation of socially necessary labor, but I feel uncomfortable with the notion of conscious robots performing labor for us.

Rev Scare wrote:Technological advance (evolutionary systems, for that matter) has proceeded at an exponential rate, and what is more, the rate of change is itself increasing exponentially. Moore's law describes a paradigm in fundamental technical approaches, but Moore's law applies only to semiconductor circuits, and it will eventually reach the climax of its S-Curve (specifically, for existing computer technology). At this point, a new paradigm shift will extend the exponential growth rate of computational devices (i.e., independent of integrated circuits); in other words, Moore's law will continue to obtain in a very general sense, although it will accelerate noticeably. The future "sixth paradigm" will likely be propelled by three-dimensional molecular computing (also known as DNA computing).

As an aside, I do not necessarily agree with all of Kurzweil's extrapolations, but I do believe that it is vital that we reflect upon his theory and its profound implications for humanity.

This is exactly what I am saying; Moore's law will cease to exist as it applies now (to existing and popular tech, such as the modern microchip) and will definitely shift and apply to the newest technology that puts the field closer to near human processing speed. It WILL collapse, but only for what it applies to now (the microchip exclusively). It will continue with the new and completely different technology that may very well come out once the semiconductor problem becomes apparent. Simply put, Moore's law as it applies to the integrated circuit will collapse.

My hope is that molecular computing or something of the same nature will be the key to further development in this field.

Celtiberian wrote:While I don't doubt that incredible technological advances are on the horizon, the notion that they are inevitable is fallacious. There are significant economic, cultural, and ecological barriers which will have to be overcome before we come anywhere near human singularity. Regarding artificial intelligence in particular, we don't even possess a full understanding of how the human mind functions. Artificial intelligence researchers, for example, have attempted structuring robotic minds around the popular massive modularity theory of human cognition (which evolutionary psychologists and certain cognitive scientists espouse), and the results have been found lacking, to put it mildly:

"...AI has failed to provide a serious account of a routine cognitive capacity, like, for example, getting across a busy street without being run over; or figuring out where to have lunch; or following an argument (to say nothing of inventing one), or playing chess; or ...etc. Simulations of cognition are, in effect, experimental tests of the architectural assumptions of the theories they embody. So far, when the architectural assumptions of Classical cog. sci. have been applied to the problems of commonsense cognition, the experiments haven't come out well; adjectives like 'appalling' spring to mind."
Fodor, J. (2005), "Reply to Steven Pinker 'So How Does The Mind Work?'" Mind & Language, 20: p. 28.

Aside from the possibility of achieving artificial intelligence, I'm undecided as to whether I even find the concept desirable. I do support the further automation of socially necessary labor, but I feel uncomfortable with the notion of conscious robots performing labor for us.

I completely understand what you mean about economic barriers and the like; my post is highly theoretical. There are many obstacles to overcome in order for this field of science to advance so far. It may never be possible. However, exploring the possibility is a worthy cause.

I do not personally think that robots will ever be able to attain exactly what makes us human. I do, however, believe that it is definitely possible for the processing speed of these robots to match and exceed our own, making them perhaps invaluable in numerous ways.

Altair wrote:It may never be possible. However, exploring the possibility is a worthy cause.

I agree.

I do not personally think that robots will ever be able to attain exactly what makes us human. I do, however, believe that it is definitely possible for the processing speed of these robots to match and exceed our own, making them perhaps invaluable in numerous ways.

I take the possibility of attaining genuine artificial intelligence quite seriously. Being that we're evolved organisms, the mechanisms by which our consciousness functions should be uncovered eventually, though I doubt it'll be any time soon (considering how rudimentary our knowledge currently is). Once known, scientists may be able to design programs capable of replicating human cognition.

The problem with prevailing theories of cognition is that some of them (e.g., the aforementioned massive modularity hypothesis) don't take into consideration the immense importance of cortical plasticity in human psychology. When researchers actually incorporate the reality of neuroplasticity into their research programs, they suddenly get results. The Marxist anthropologist Chris Knight described one such study, in which an artificial language was developed:

"Why not build whole communities of robots capable of moving and interacting in real physical environments?

It seemed worth a try. [Luc] Steels dreamed of sophisticated robots that could learn from their own experience. No grammatical constraints would be specified in advance. Instead, Steels envisaged robots capable of proceeding by trial and error. As they sought to communicate, they would remember what worked and forget what didn't. In this way, Steels reasoned, a system resembling language might spontaneously evolve. If this were to happen, the emergent grammars might display all sorts of unexpected features—details so complex that no human engineer could possibly have fixed them in advance.

Steels is today perhaps the most celebrated robot-maker in the world. His machines have limbs, sense organs and sophisticated brains; they move around and navigate in pursuit of various objectives; finally, they co-operate and communicate in what would appear to be ingenious and creative ways. The minds of these robots are not hard-wired for language in advance. Rather, they actively configure one another as they socially interact, recruiting for linguistic purposes whatever cognitive capacities and procedures are available. Naturally, the languages which emerge are not quite as complex as, say, English or Warlbiri. But as they evolve, they do embody those principles of 'discrete infinity' which, according to Chomsky, constitute the generative source of every human language."
Interview with Christ Knight. "Human Revolution and the Origins of Language" (2006).

The obvious concern with developing conscious machines is that they could theoretically experience feelings of pain and/or exploitation. It would therefore be unethical to use them as we would use a tractor, for instance. In which case, why even waste valuable resources developing them? I do believe that advanced technologies should be developed, and I agree that they could prove to be very liberating for the humanity. But I think it would be advisable not to invest a great deal of energy into striving to attain what some theorists call "strong AI."

Celtiberian wrote:I take the possibility of attaining genuine artificial intelligence quite seriously. Being that we're evolved organisms, the mechanisms by which our consciousness functions should be uncovered eventually, though I doubt it'll be any time soon (considering how rudimentary our knowledge currently is). Once known, scientists may be able to design programs capable of replicating human cognition.

The problem with prevailing theories of cognition is that some of them (e.g., the aforementioned massive modularity hypothesis) don't take into consideration the immense importance of cortical plasticity in human psychology. When researchers actually incorporate the reality of neuroplasticity into their research programs, they suddenly get results. The Marxist anthropologist Chris Knight described one such study, in which an artificial language was developed:

"Why not build whole communities of robots capable of moving and interacting in real physical environments?

It seemed worth a try. [Luc] Steels dreamed of sophisticated robots that could learn from their own experience. No grammatical constraints would be specified in advance. Instead, Steels envisaged robots capable of proceeding by trial and error. As they sought to communicate, they would remember what worked and forget what didn't. In this way, Steels reasoned, a system resembling language might spontaneously evolve. If this were to happen, the emergent grammars might display all sorts of unexpected features—details so complex that no human engineer could possibly have fixed them in advance.

Steels is today perhaps the most celebrated robot-maker in the world. His machines have limbs, sense organs and sophisticated brains; they move around and navigate in pursuit of various objectives; finally, they co-operate and communicate in what would appear to be ingenious and creative ways. The minds of these robots are not hard-wired for language in advance. Rather, they actively configure one another as they socially interact, recruiting for linguistic purposes whatever cognitive capacities and procedures are available. Naturally, the languages which emerge are not quite as complex as, say, English or Warlbiri. But as they evolve, they do embody those principles of 'discrete infinity' which, according to Chomsky, constitute the generative source of every human language."
Interview with Christ Knight. "Human Revolution and the Origins of Language" (2006).

The obvious concern with developing conscious machines is that they could theoretically experience feelings of pain and/or exploitation. It would therefore be unethical to use them as we would use a tractor, for instance. In which case, why even waste valuable resources developing them? I do believe that advanced technologies should be developed, and I agree that they could prove to be very liberating for the humanity. But I think it would be advisable not to invest a great deal of energy into striving to attain what some theorists call "strong AI."

Exactly. The human brain itself is far from being completely understood; how then can we hope to understand how to successfully give humanesque knowledge to a robot when we do not fully understand the brain ourselves?

Indeed, our ability to adapt and change ourselves in accordance with life's events is invaluable. Implanting that state of consciousness into robots could prove to be dangerous, especially if exploitation of these human-robot hybrids were to occur. The exploitation itself is disturbing...first we exploit people, and then robots that can feel and understand as we do? Frankly, I find the notion to be sickening. If such an advanced AI were to be invented, I hope it would not be by the hands of capitalists. People can die in an instant but robots can be much more durable...a never ending supply of wage slaves. Think of what could happen if used by our enemies. It is stomach churning.

Putting this issue above current concerns would be a mistake. We have far more important issues that are knocking on the door as we speak. Only after our society has undergone a revolution would I even suggest toying with this type of sophisticated AI.

Altair wrote:Implanting that state of consciousness into robots could prove to be dangerous, especially if exploitation of these human-robot hybrids were to occur. The exploitation itself is disturbing...first we exploit people, and then robots that can feel and understand as we do? Frankly, I find the notion to be sickening.

As do I. The problem is, capitalists invest in technology exclusively for issues related to production and profit, while governments do so for military purposes. From a socialist perspective, both uses are completely unethical if conscious beings are exploited. It's conceivable that, within a socialist mode of production, scientists may be able to persuade society that investing in AI research is worthwhile for more honorable reasons, and therefore succeed in acquiring the necessary resources to do so. I certainly wouldn't be opposed the development of, say, one artificially intelligent robot solely for matters of scientific inquiry.

If such an advanced AI were to be invented, I hope it would not be by the hands of capitalists. People can die in an instant but robots can be much more durable...a never ending supply of wage slaves. Think of what could happen if used by our enemies. It is stomach churning.

Strong AI and genetic engineering, if developed under current bourgeois social relations, would be absolutely disastrous for a whole host of reasons. It is therefore the task of radical organizations to do their utmost to inform the public about these issues and ensure that the bourgeoisie are not allowed to use these technologies to their advantage.

Putting this issue above current concerns would be a mistake. We have far more important issues that are knocking on the door as we speak. Only after our society has undergone a revolution would I even suggest toying with this type of sophisticated AI.

Absolutely.