International regulation of autonomous weapon systems (AWS) is increasingly conceived as an exercise in risk management. This requires a shared approach for assessing the risks of AWS. This paper presents a structured approach to risk assessment and regulation for AWS, adapting a qualitative framework inspired by the Intergovernmental Panel on Climate Change (IPCC). It examines the interactions among key risk factors—determinants, drivers, and types—to evaluate the risk magnitude of AWS and establish risk tolerance thresholds through a risk matrix informed by (...) background knowledge of event likelihood and severity. Further, it proposes a methodology to assess community risk appetite, emphasizing that such assessments and resulting tolerance levels should be determined through deliberation in a multistakeholder forum. The paper highlights the complexities of applying risk-based regulations to AWS internationally, particularly the challenge of defining a global community for risk assessment and regulatory legitimization. (shrink)

This article provides a methodology for the interpretation of AI ethics principles to specify ethical criteria for the development and deployment of AI systems in high-risk domains. The methodology consists of a three-step process deployed by an independent, multi-stakeholder ethics board to: (1) identify the appropriate level of abstraction for modelling the AI lifecycle; (2) interpret prescribed principles to extract specific requirements to be met at each step of the AI lifecycle; and (3) define the criteria to inform purpose- and (...) context-specific balancing of the principles. The methodology presented in this article is designed to be agile, adaptable, and replicable, and when used as part of a pro-ethical institutional culture, will help to foster the ethical design, development, and deployment of AI systems. The application of the methodology is illustrated through reference to the UK Ministry of Defence AI ethics principles. (shrink)

The question of human control is a key concern in autonomous military systems debates. Our research qualitatively and quantitatively investigates values and concerns of the general public, as they relate to autonomous military systems, with particular attention to the value of human control. Using participatory value evaluation (PVE), we consulted 1980 Australians about which values matter in relation to two specific technologies: an autonomous minesweeping submarine and an autonomous drone that can drop bombs. Based on value sensitive design, participants were (...) tasked to enhance the systems with design features that can realize values. A restriction (limited budget) in each design task forced participants to make trade-offs between design options and the values that these options realize. Our results suggest that the ‘general public’ has diverse and nuanced stances on the question of human control over autonomous military systems. A third of participants that is opposed to autonomous military systems when asked directly, selected different combinations of design features realizing varying degrees of human control. Several contextual factors, technology-specific concerns, and certain values seemed to explain these different choices. Our research shows that a focus on human control might overlook other important values that the general public is concerned about, such as system reliability, verifiability, and retrievability. (shrink)