In March 2020, as Covid-19 locked down the world, what may have been the first autonomous drone attacks in history were taking place on a largely unwatched battlefield in Libya.
According to a UN report published in March this year, Libyan forces loyal to the Government of National Accord (GNA) used Turkish-made STM Kargu-2 drones to hunt down units loyal to former Libyan Field Marshall Khalifa Haftar.
The report stated that the four-rotor drones were programmed in "autonomous mode" to attack fleeing logistics convoys and other vehicles automatically, without further human intervention.
If correct, that would represent the first documented such incident on a battlefield – a development that has long been predicted and feared by military and human rights experts alike. While drones have been a feature of the battlefield for years, they have still required a human being to operate the "kill switch".
That such a step may have taken place unheralded and largely unnoticed, however, should not be a surprise. The last two decades have seen a mass proliferation, downsizing and democratisation of technology once the preserve of the most powerful states. Innovation is now much cheaper, and those willing to bend rules can find advantage.
Some campaigners have long called for a prohibition on autonomous killer drones. They had argued this should be done before such technology became a reality, but it may already be too late. Whether or not the March 2020 Libya strike was the first autonomous drone attack, it is unlikely to last – and there may well have been more since then.