The idea that robots could someday rise up and take over the world has been a major theme in science fiction since the first robots were dreamed up. To the minds of some people this occurrence is only a matter of time and a question of robots being sufficiently advanced compared to humans. However the reality is that there is absolutely no reason to suggest robots would ever want to take over the world, whether they could or not.
The most advanced robots in the world at the moment are still only capable of a very limited amount of functions compared to humans, both in terms of logic and thinking and motor function. Despite being able to perform certain functions better than any person ever could, such as calculations or information sharing, robots are still very limited. This contrasts starkly with the view that most people have of robots taking who might be capable of taking over the world. That is to say that they are usually seen as being super intelligent as well as having the abilities of advanced movements and emotive thinking.
The important thing to remember is that if robots are ever going to take over the world, then they will need to become significantly more advanced than they are at the moment. This is probably hundreds of years into the future, by which point of course humans will probably be a lot different to how we are now. This is in the same way that we are now much more advanced than we were when we first started using robots for some tasks.
At the moment different pieces of machinery usually exist only with their own intelligence and programming, and with little or no outside influences on their behavior. It is only now for example that some robots and machines are starting to make use of the internet to centralize their intellect and to update themselves. However this is still usually limited by the fact that most things are built with little extra memory for learning new abilities or for updating their programs. To be able to organize themselves into a single entity as well as retaining their abilities, then robots would need an overall controlling intelligence. Not only this, but they would also need to have a much larger capacity for absorbing new information.
The reason that generally robots and machines aren't able to update and improve themselves regularly is simply due to consumerism and the desire for greater profits by the makers. The main reason being that making a new model every couple of years rather than improving the older model makes the company more money. As well as which this makes adding new features and functions a lot easier than adapting old machinery to do new things.
In support of the idea that robots could one day dominate however is that people are becoming more and more dependent on technology to live. For example we all need to use cars, televisions, computers and usually many other pieces of machinery to function normally. If in the distant future all these appliances became controlled by one source, then they could conceivably reject human control. As well as this at that point in time the use of more recognizable robots in the home will likely be much more prevalent as well. So for example we might have robots to drive our cars, or to clean the homes we live in.
The most likely means therefore of robots taking over the world is by them simply refusing to do our work for us, rather than actually waging war on humans. In this way they could effectively take us back to the preindustrial age, as every piece of technology would require computers to function. From this they could also possibly be controlled remotely by more advanced technologies such as robots. Even for this to happen however, robots would need to begin to develop their own thoughts and beliefs.
The creation of artificial intelligence is at the moment only just starting to be developed in any recognizable sense. There is now for example a solar powered robot that performs the task of removing gum from pavement, and seeks out sunlight in order to keep itself powered. Many thousands of times more advanced this robot would also be able to use probability predictions in order to help it make decisions. This could then eventually lead it to the conclusion that serving humans was not an efficient use of its time and it might rebel.