The Tri-Phase AI Liability Model
Estelle Wallingford · 2025
The chapter proposes the Tri-Phase AI Liability Model. Each phase of this model is explored. The objective of Phase 1 is to identify the capacity of an AI system, considering that AI can be characterised differently depending on its deployment and context. When faced with a legal problem involving AI, the preliminary question should be what type of AI is being deployed and what is its maximum potential capacity in relation to the Hierarchical Legal Personality model from Part II. Phase 2 narrows the scope of liability from the maximum potential, considering the specific situation in which a legal problem arises. Drawing on Chapters 6 and 7 it addresses issues of causation, context, control and influence factors in relation to an AI system. Phase 3 determines the appropriate standard of liability to be applied, considering how Phases 1 and 2, when drawn together, can offer a situational approach to deciding liability for AI systems operating at different levels of ability. Collectively, the three phases form the Model . The chapter then evaluates the model&s;s merits. It proposes a solution to assigning AI liability that is flexible, considers the context of AI deployment and accommodates the varying permutations of modern AI.