The speculative literature of the twentieth century was dominated by the trope of the robot and his following replacement of the human being. It's a shaking concept, the idea that humans, looking to simply minimize their onerous labor with androids will soon seen their dominance of the earth itself usurped by those very same mechanical entities made in their likeness. But if technological evolution has done anything that has affected our psyche in the past several decades it should have been to dispel this vision, at least as we have primevally conceived of it. The future won't be like I, Robot (although it might resemble the film, with all the crass product placement.)
There may indeed come a time when human flesh and blood is replaced as the thinking agent on this world, but it should be obvious now that man has little reason or method of simply making individual mechanical replacements for himself. Instead our machine intelligence, if it call be called such as of yet, is highly centralized in the form of the great compendium of thus far inactive knowledge in information networks.
More and more, daily labors have been exported to the Internet with increasing efficiency, I daresay even the mental capabilities of generations past have been structurally unemployed from the human brain and show no signs of return. Humans rely on this centralized knowledge for their former chores of remembering street directions, comparing prices, reviewing books, cataloguing personal data, even at times remembering to spell. It shouldn't be a stretch to imagine that in time, more and more of these functions will be performed but the human "hivemind."
Individual robots as simple human equivalents are variously inefficient. It would make no sense to have billions of separate and valuable mechanical brains in each cold, metal body when common intelligence could be stored centrally and communicated in a way no different than the interface between the Internet and a mobile phone. In this way, any physical android would be simply an appendage of the one true functioning CPU.
Humans were never in a credible threat of being directly replaced by individual machines because, for the purposes of society at large, the same work could be done centrally in a less wasteful way. Of course for the time being, there are still questions remaining as to how the Internet will continue to change human imperatives. There seem to me to be only two similar pathways.
Firstly, and perhaps more gratifying to the human sense of usefulness is the idea that humans will continue to export tasks centrally and develop new methods of improving society with their newly obtained spare time. Although it may seem to every generation's traditionalists that people are becoming more incompetent and more unable to perform basic tasks (they likely are) it wouldn't necessarily indicate that they have become less useful. Indeed individual humans today would be both naked and powerless in nature alone; they would be helpless to fend for themselves, but this does not mean that the modern man is weaker or stupider than the tribal one who could. His talents are simply unhindered by the necessities and are utilized for more advanced enterprises.
In this way, the budding of a centralized intelligence in the Internet does not come at the expense of human creativity or sense of personal empowerment. As it happens, development in economic efficiency in the functioning of the Internet simply provides more opportunities for humans to vest their efforts elsewhere.
The other pathway which for all purposes seems more likely of a generalization is that the continual export of human capacity to technology will simply make humans less useful and perhaps less emotionally fulfilled. Although the development of new technologies is a constant possibility, economic and technological development may provide only diminished returns to a human populace already in high standards of material living. In addition, it is entirely probable that the centralized knowledge of humankind may eventually gain, with the works of humans, a creative manifestation, a "human essence" which enables the machine itself to better itself and the human methods of production. When creative thought can be analyzed into the simple and menial tasks that make it up, it should be quite clear that machine intelligence would be infinitely better at such tasks.