As artificial intelligence (AI) becomes increasingly integrated into military operations, the ethical landscape of warfare is undergoing a profound transformation. A recent study by Susannah Kate Devitt explores the complex issue of moral responsibility for civilian harms in human-AI military teams, raising critical questions about accountability in modern conflict.
The study identifies three key categories of soldiers: “bad apples,” “mad apples,” and “cooked apples.” Bad apples are those who commit war crimes, while mad apples are soldiers who, due to extreme stress or psychological trauma, may not be fully responsible for their actions. Cooked apples, however, represent a new and troubling category: soldiers who are placed in untenable decision-making environments due to the increasing reliance on AI in warfare. This reliance can lead to a detachment from the moral consequences of their actions, potentially resulting in extreme moral witnessing, becoming moral crumple zones, or suffering moral injury.
The integration of AI into military decision-making processes raises significant ethical concerns. As AI systems take on more responsibility for targeting and strategic decisions, the lines of accountability become blurred. Soldiers may find themselves operating within systems where their moral agency is compromised, leading to a detachment from the ethical implications of their actions. This detachment can manifest in various ways, including extreme moral witnessing, where soldiers become desensitized to the consequences of their actions, or becoming moral crumble zones, where they absorb the moral blame for decisions made by AI systems.
Devitt’s research highlights the need for new mechanisms to map out conditions for moral responsibility in human-AI teams. One proposed solution is the implementation of new decision responsibility prompts within a critical decision method in cognitive task analysis. These prompts would help soldiers and AI systems navigate the ethical complexities of their roles, ensuring that moral responsibility is clearly attributed and understood.
Additionally, the study suggests applying an AI workplace health and safety framework to identify cognitive and psychological risks relevant to attributions of moral responsibility in targeting decisions. This framework would help militaries design human-centred AI systems that prioritize ethical considerations and support the well-being of soldiers. By addressing these risks proactively, militaries can create environments where soldiers are better equipped to handle the moral challenges of modern warfare.
The study also underscores the importance of acknowledging military ethics, human factors, and AI work to date, as well as critical case studies. By learning from past experiences and current research, militaries can develop more robust ethical guidelines and protocols for the use of AI in military operations. This approach ensures that the integration of AI is not only technologically advanced but also ethically sound.
In conclusion, as AI continues to reshape the landscape of military operations, the ethical implications of human-AI collaboration must be carefully considered. Devitt’s research provides a crucial framework for understanding and addressing the moral responsibilities of soldiers in an era of AI-driven warfare. By implementing new decision-making prompts and workplace health and safety frameworks, militaries can ensure that their use of AI is both effective and ethically responsible, ultimately safeguarding the moral integrity of their soldiers and the civilians affected by conflict. Read the original research paper here.