On Ethical Autonomous Weapons Systems

Something that’s been blinking on and off the radar in the last couple of years is the idea of “ethics” in weapons systems. There’s an interesting article here, that argues fairly cogently that it’s only a matter of time before we go from “humans must be present (and sober) in robotic cars” to “humans are not allowed to drive cars at all”… and that this scenario might also apply to weapons-systems.

Which is something (in a funny sort of way) I have some sympathy for, because most of the really evil shit that happens in war (including wars being started in the first place) comes out of bent human emotions and misunderstandings – failures of perspective. Robots would not be prone to Stamford Prison Experiment syndrome. Shots would not be fired in anger.

What “ethical killing systems” would be up against of course is 100 years of popular culture, in which humanity is attacked by its scientifically-spawned progeny. Skynet-Syndrome… so instead it’s highly likely that what we’ll see is a proliferations of mercenary/privatised devices that can operate outside the “laws” (or at least the public accountability) that governments feel they need to operate within… although American wars (ie: occupations and extra-judicial killings) are already happening outside the laws that the rest of the planet are expected to adhere to. I’m not sure how sustainable that is within a highly networked world. The biggest city of twitter users is (for example) Jakarta.


When people start talking among themselves, things change. That Jaron Lanier recently opined that “laser-beams of cruelty” were likely to arise from the web… playing the “pogrom in personal history” card, but I think that he’s completely mis-attributing causality to “people talking among themselves”… when what actually causes lynch-mobbery is sustained feelings of frustration, alienation and powerlessness. Austerity imposed from on-high in other words. It’s not The Internet that’s causing the Golden Dawn to rise in Greece, it’s the state creating a vacuum of competence… which is happening to a greater or lesser extent in every government on earth, because every government on earth (with about 4 exceptions) are attempting to serve the needs of an increasingly illegitimate globalised aristocracy, in an environment where secrecy is becoming increasingly difficult.

But it’s not the failure of secrecy that’s the problem. It’s the corruption that’s the problem.

But I digress.

Back to ethical weapons systems… Humans are a chest-beating species. We’d far prefer to beat chests and scare each other than actually fight – which has been a real problem for armies in the past – where a huge percentage of shots fired at the enemy were deliberately fired high… in fact in the American Civil War, muskets were found rammed full of shot – because people were reloading without firing. The Nato countries figured this out – so have trained their soldiers to reflexively kill… during the Falklands War the kill rate of UK soldiers was massively higher than Argentinian… but they didn’t train soldiers to deal with the aftermath, and now more Falklands soldiers have committed suicide than were actually killed by their supposed enemies.

It’s about chest-beating and brinkmanship… which is (in humans) about being able to horrify with violence. This is another major reason why human wars get so ugly… violence is theatre of horror.

So… if it’s possible to make robots that are “ethical”, then it’s also possible to make them unethical. Deliberately demonic. The Nuclear Option. And if we can do this, we will.