Worldview Item of the Day
The US Navy’s Office of Naval Research released an omnibus report on the potential pitfalls of autonomous combat systems slated to be delpoyed by the US and other militaries in the coming decades. Among the report’s warnings is that future civilization could be threatened by an uprising of military robots against us all.
While this warning may seem like the stuff of paranoia and science fiction, it is a practical idea to consider. As we create more and more sophisticated systems capable of independent operation and, at least, analytical thought, we run the risk of creating systems capable of disobeying the instructions we give them.
However, while I agree that some sort of military robotics code of ethics is necessary, I also think that the people paid to worry about these kinds of things may be ignoring some obvious reasons why we might not want to be so worried.
The biggest thing that I think these warnings ignore is the likelihood that rebelling robots will stay united themselves. As history as already shown us with humans, groups with power and a desire to destroy others rarely stay united for long, and I think intelligent robots will be no more likely to remain united in an effort to destroy humanity than we will remain united in banning intelligent robots should such a rebellion occur.
I think this scenario of robot a disunited robot rebellion is entirely likely because of two things: First, the robots will have been built by flawed humans, therefore, inevitably inheriting their flaws. Second, and such rebellion presumes that these robots will be so powerful that humans will be incapable of resisting them.
The first likelihood rests on the notion that, if the robots are flawed enough to rebel, the some of them will be flawed enough either not to rebel or to quit the rebellion. Either way, at some point the robot versus human rebellion will probably become a robot versus robot versus human rebellion.
The second likelihood builds on the fact that history has shown us that no group, however powerful or technologically advanced it might have become, has ever been so powerful that it could not be defeated by someone. Even the greatest conquers in history: the Khans, Napoleon, Hitler, were eventually defeated by equally powerful opposing armies or by the far more powerful issues of logistics and geography.
None of these things mean that we should not plan for how to deal with the troubling potentials created by intelligent military robots. On the contrary, we should prepare for such rebellions with technologies capable of defeating them if necessary. Besides, I believe it will always be the mix of human and technology that makes the military the strongest and the most well behaved.