The BIG Argument for AI Safety Cases
Ibrahim Habli, Richard Hawkins, Colin Paterson, Philippa Ryan, Yan Jia, Mark Sujan, John McDermid
TL;DR
The paper introduces the Balanced, Integrated and Grounded (BIG) argument as a whole-system safety-case framework for AI across varying capabilities, emphasizing a balance between safety and ethics, integration of social, ethical, and technical dimensions, and grounding in established safety norms. It structures safety arguments using a modular, GSN-based approach comprising four sub-arguments (Ethical, System, Purpose-specific ML, and General-Purpose GPAI) and connects them with patterns such as PRAISE, SACE, and AMLAS to address both context-sensitive risks and frontier-model capabilities. The authors illustrate how to justify safety through evidence across the ML lifecycle, data management practices, robustness testing, and real-world case studies, including sepsis treatment and wildfire-detection examples, while highlighting the need for independent audits and red-teaming in frontier AI. They argue for traceability, case-study development, and dynamic updating of safety cases to keep pace with AI's rapid evolution, and stress the importance of sociotechnical considerations to avoid narrowing safety to technical safeguards alone. Overall, the BIG argument offers a comprehensive, adaptable blueprint to improve transparency, accountability, and justifiability in AI safety across domains and deployment contexts.
Abstract
We present our Balanced, Integrated and Grounded (BIG) argument for assuring the safety of AI systems. The BIG argument adopts a whole-system approach to constructing a safety case for AI systems of varying capability, autonomy and criticality. Firstly, it is balanced by addressing safety alongside other critical ethical issues such as privacy and equity, acknowledging complexities and trade-offs in the broader societal impact of AI. Secondly, it is integrated by bringing together the social, ethical and technical aspects of safety assurance in a way that is traceable and accountable. Thirdly, it is grounded in long-established safety norms and practices, such as being sensitive to context and maintaining risk proportionality. Whether the AI capability is narrow and constrained or general-purpose and powered by a frontier or foundational model, the BIG argument insists on a systematic treatment of safety. Further, it places a particular focus on the novel hazardous behaviours emerging from the advanced capabilities of frontier AI models and the open contexts in which they are rapidly being deployed. These complex issues are considered within a wider AI safety case, approaching assurance from both technical and sociotechnical perspectives. Examples illustrating the use of the BIG argument are provided throughout the paper.