Much has been made about the use of ‘autonomous’ drones. Many writers confuse autonomy with ‘pilotless’.
But clearly, a fully autonomous drone is one that after launch carries out a task without further interaction with the human controller, based on a pre programmed set of instructions, or decision making rules.
Thus, autonomous drones might survey a target and make a decision whether or not to attack, without further human guidance. This level of accuracy can’t be envisioned now.
Why is this dangerous? First of all it implies that the technology permits the drone to accurately follow the instructions and not make a mistake. This level of accuracy can’t be envisioned now.
More disturbingly, the autonomous drone will almost certainly be stealthy and surrepticious and will thus provide a high degree of deniability and anonymity to the controllers. Furthermore, this ‘killing at a distance’ reduces the moral hazard of the controllers. A controller assured of relative safety from identification or prosecution will programme an autonomous drone differently than one who has to defend his/her actions.
Because drones will be cheap, this means that an entirely new venue for warfare has been opened up. It also opens up a new venue for action by non-state actors, quasi state actors, police, security and intelligence organisations, etc.
Use of drones by these organisations will be difficult to monitor because autonomous drones of the future need not cost a lot, and therefore won’t be large line items in budgets. Nor need they be a large physical presence.
There are already many autonomous weapon systems. Samsung produces an autonomous sentry system capable of spraying a zone with machine gun fire if it detects movement. (see categories list right). Similar systems have been proposed for Israeli border areas.