One of the many areas in which automation is steadily advancing is in weapons systems. Advances in machine learning systems that can parse photos and video, recognize faces, maneuver in complex 3-dimensional spaces, etc., can in principle allow new weapons systems that operate largely or wholly without human guidance.

As described here, such weapons raise a number of both strategic and ethical questions involving the threshold of conflict, arms races, and who (or what) chooses to take human lives. Several campaigns have arisen calling for an international ban on lethal autonomous weapons.

One major concern raised by such campaigns, articulated for example in this open letter, is that an arms race in autonomous weapons could lead to cheap, widely available, highly effective weapons that could be used for political purposes including suppression of dissent or assassinations. For example, a swarm of tiny drones with facial recognition systems could seek out particular individuals (or groups) and kill them with toxins or small close-range explosives.

Will a credible media report indicate that an autonomous weapon system has been used to kill a political figure by start of 2025?

Positive resolution requires that:

• the figure killed is in a leadership role of a political group – either a government or other organization built around political ends, and

• the target is identified by the autonomous system itself, according to some criteria, rather than by other means of surveillance (which may be used to localize the target but not select the target out of, for example, nearby people), and

• no other "unintended" people are significantly harmed in the attack.