International relations professors have debated the implications of zombies for world politics. But no one, with the exception of Joseph Fouche (always ready to think the unthinkable) has pondered the international relations of a robot uprising. Since the task of pondering the robot is a vast task (to be tackled with frequent lolcat and caffeine breaks), I think we should start by looking at a few case studies of different tapes of human-robot conflict in science fiction.
This is the most obvious and recognizable human-robot conflict. It begins with a robot nuclear coup de main that fails to annihilate humanity, thus showing that even robots with control of the world's nuclear arsenal find properly executing a war of annihilation with standoff-based firepower difficult. The conflict eventually turns into a grinding war of attrition. The humans are globally dispersed, but have a central mobile C2 node (the submarine in Terminator Salvation).
While humans can use compound warfare tactics (they have mechanized air and ground forces), command and control over the human resistance forces is difficult, especially because the legitimacy of the top leadership is in doubt. It is too small to exercise a span of control over the global human forces. Moreover, the humans are massively out-numbered by self-replicating machines able to exercise overwhelming combat power.
Since this is purely a war of extermination, the robots will not rest until they find and destroy every last bit of human civilization. There is no viable human strategies exercised by the human leadership that could plausibly explain victory except magic bullets i.e keeping John Connor alive so he can save the world in some unexplained way. This is understandable, as the most likely human intellectual reaction to a worldwide nuclear catastrophe and the threat of extinction by murderous robots would probably be panic, disorganization, and pseudo-religious cults of personality.
2. The Matrix
What is interesting about The Matrix is the diversity of types of warfighting that occur. We start with the prequel supplement, the Animatrix. Diplomatic miscalculations by desperate humans seeking to destroy "01"--the economically and militarily superior robot nation--before it could become hegemonic over the international system led first to conventional warfare. The humans tried and failed to annihilate the robots with a preemptive strike with large amounts of nuclear weapons and conventional maneuver forces. This fails to destroy enough robots and enough are reproduced to continue the fight.
The robots then launch rapid decisive operations that leverage their superior technology and overwhelming numbers, successfully waging a campaign of strategic annihilation. After a last-ditch attempt to deny the robots solar power, the human army is then completely destroyed. Any humans left alive are forced into slavery. A low-level insurgency, operating from a secure base area, predictably ensues in which the humans carry out a "Focoist" strategy of urban guerrilla warfare against robots and their human servants--conducted entirely inside life-like virtual worlds.
The Matrix robots have similar advantages to the robots in Terminator, except this time the humans have less resources and are all concentrated in one vulnerable location. While the robots are smarter than Terminator's bots (Agent Smith being a good example), the humans do not even have the mechanized forces available to John Conner's army in Terminator Salvation. The human leadership is divided and has no coherent plan to defeat the machines. It's only another "magic bullet" solution (Neo as The One) rooted in a pseudo-religious cult of personality that allows humankind to survive.
3. Blade Runner
There are other examples (Crispin Burke is probably raising his hand right about now, asking "What about Battlestar Galactica???") but for the purposes of diversity it's important to pick one that features a decisive human advantage. In Blade Runner, robots are a "law enforcement problem." Humankind depends on them for a variety of functions (off-planet labor being one), but faces the problem of properly controlling them. These robots are strong, homicidal, and mentally unstable--with a history of escape. Some humans also illegally use robots for various unregulated tasks. Even more problematic is the fact that only a small portion of humans know how to tell robots and humans apart.
The robots are not, however, waging an insurgency. Instead, they're essentially committing small-time acts of criminal violence that fall short of the definition of "war," although there are scenarios implicit in the universe that might analogized to early 20th-century anarchist terrorism. Most importantly, the structural limitations of their limited lifespans prevent them from becoming a strategic threat to human civilization.
That being said, the threat of their homicidal violence and crime is serious enough for humankind to rely on a corps of Old West-style bountyhunters ("Blade Runners") to hunt down and eliminate the robots. The human strategy is thus one of risk management. The threat of robot violence is not high enough to justify ceasing to manufacture robots or implementing harsher security measures, so humans instead have constructed sophisticated privatized robot-hunting agencies authorized to terminate any rogue robots with extreme prejudice.
3. Overall Lessons
Terminator and The Matrix both suggest that a terrestrial war against robots offers no real hope for humanity. The most likely scenario for humanity in these films is either a doomed Fabian strategy or a similarly pointless campaign of human urban guerrilla warfare. Pseudo-religious cults and divided political leadership are both strong features of human organizations in the film's robot-dominated world.
In contrast, Blade Runner suggests that it is possible to live in a world where robot violence, albeit sporadic and uncoordinated, is a feature of everyday life. Keeping the level of robot violence and crime under an acceptable level requires specialized talents who can identify and neutralize out-of-control machines. Perhaps the problem might be better managed, however, if robot manhunting was centralized rather than ad hoc.