Autonomous weapons systems (AWS), sometimes referred to as “killer robots”, are receiving evermore attention, both in public discourse as well as by scholars and policymakers. Much of this interest is connected with emerging ethical and legal problems linked to increasing autonomy in weapons systems, but there is a general underappreciation for the ways in which existing law might impact on these new technologies. In this paper, we argue that as AWS become more sophisticated and increasingly more capable than flesh-and-blood soldiers, it will increasingly be the case that such soldiers are “in the power” of those AWS which fight against them. This implies that such soldiers ought to be considered hors de combat, and not targeted. In arguing for this point, we draw out a broader conclusion regarding hors de combat status, namely that it must be viewed contextually, with close reference to the capabilities of combatants on both sides of any discreet engagement. Given this point, and the fact that AWS may come in many shapes and sizes, and can be made for many different missions, we argue that each particular AWS will likely need its own standard for when enemy soldiers are deemed hors de combat. We conclude by examining how these nuanced views of hors de combat status might impact on meaningful human control of AWS.

Disorders of swallowing often lead to pneumonia when material enters the airways (aspiration). Flexible Endoscopic Evaluation of Swallowing (FEES) plays a key role in the diagnostics of aspiration but is prone to human errors. An AI-based tool could facilitate this process. Recent non-endoscopic/non-radiologic attempts to detect aspiration using machine-learning approaches have led to unsatisfying accuracy and show black box characteristics. Hence, for clinical users it is hard to trust in these model decisions. Our aim is to introduce an explainable artificial intelligence (XAI) approach to detect aspiration in FEES. Our approach is to teach the AI about the relevant anatomical structures like the vocal cords and the glottis based on 92 annotated FEES videos. Simultaneously, it is trained to detect bolus that passes the glottis and becomes aspirated. During testing, the AI successfully recognized glottis and vocal cords, but could not yet achieve satisfying aspiration detection quality. Albeit detection performance has to be optimized, our architecture results in a final model that explains its assessment by locating meaningful frames with relevant aspiration events and by highlighting the suspected bolus. In contrast to comparable AI tools, our framework is verifiable, interpretable and therefor accountable for clinical users.