christophe malvasio Thanks for your comment. Humans may indeed retain the ‘destroy button’ to fire weapons from unmanned military vehicles. But, it is perfectly possible that due to an increasing speed of warfare humans may not be able to keep up with machines, and therefore the kill-decision could be delegated to autonomous systems. Although it would be morally ideal to keep all lethal decisions with humans, in a future high-intensity conflict between high-tech states, the delegation issue could determine who wins the conflict. So the use of autonomous weapon systems with lethal decision-making delegate to them is certainly a possibility.
In terms of governments showing unlawful incidents as an accident, international human rights law guarantees an effective remedy to any breaches of human rights under Art.2(3) of the International Covenant on Civil and Political Rights. This ‘remedy’ includes an effective investigation into breaches of human rights. Therefore, any unlawful killing by an autonomous system should have an effective investigation into it, which would preclude governments from simply suggesting accidents as a way to get around malfunctioning, or poorly programmed AWS. This does, however, require states to abide by their human rights obligations.
Cyrus Epler thank you for your comment. You can see the list of countries which have called for a ban here. None of them has a particularly advanced defence or tech industry. So I wouldn’t think that they are stating their opposition to AWS in order to obfuscate secret military research. The countries who have stated they do not support a ban are generally those who have carried out research on unmanned military vehicles, and the use of autonomy in military contexts. These countries are fairly open about performing this research, although obviously do not share detailed results. They are also fairly open about their production and possession of some systems with autonomy. You can access a database of weapons with autonomy here. I hope this helps.