New drone has no pilot anywhere, so who’s accountable?
The Navy is testing an autonomous plane that will land on an aircraft carrier. The prospect of heavily armed aircraft screaming through the skies without direct human control is unnerving to many.
http://www.latimes.com/business/la-fi-auto-drone-20120126,0,740306.story
While it is a bit premature to be serious about self-aware robots going on killing sprees, I don’t think it is too premature to think about how these things should be addressed in future planning. While it is _possible_ to build a system that is immune to external takeover (I am not sure I have the confidence that our military could do such a thing), there is absolutely no way to guard against an internal threat taking over such a system. By putting more and more power into fewer and fewer hands it becomes easier and easier for adversaries to take control. The upside to a chaotic system where so damn many people have to be involved in even the most trivial of decisions is that it is very rare for decisions to happen in a vacuum. Turn that dynamic on its head and you can have dangerous situations where a Strangelovian situation arises where one man (or woman, let us not be sexist here) can unilaterally go to war. Unlike scenarios like Fail Safe where there are humans in the loop (unfortunate for the movies mentioned, those humans carried out their duties with robot-like precision, to the detriment of two huge cities and millions of people), once humans have been removed from the decision making loop (or so greatly reduced that one or just a handful of people can implement decisions) then there will be no hope of any sort of callback once a fateful decision has been made. Unless, of course, you have some sort of universal kill switch, which, of course, makes your whole force acutely vulnerable (which I am sure would never make it into production).
Anyway, drones are all operated by humans, just have the same autonomy as a today’s autopilot found on traditional airliners. The plane is not making any sort of decisions, certainly nothing about mission, weapons activation, targets, etc. What is described above is the first steps toward a truly autonomous system that will have to have at least minimal decision making capability. I see this as inevitable due to nothing more that cost cutting and increasing efficiency in our military. Unfortunately, as stupid as it seems to go down this route, it is just as stupid to not to (the only rational path is to do away with the military on this scale all together, not something I think is at all feasible in our current political climate). Just like in the Terminator movies, Armageddon hasn’t been avoided, just delayed.
One thought on “Skynet in its infancy”
Comments are closed.