Responsibility in Extensive Form Games
This work addresses a foundational issue in AI and philosophy for multi-agent systems, though it is incremental as it builds on prior modalities.
The paper tackles the problem of defining 'seeing-to-it' responsibility in sequential multi-agent settings, where existing definitions fail to capture intuitive responsibility. It proposes a new amalgamated definition, shows it is distinct from counterfactual responsibility, and demonstrates that a responsibility gap can be resolved by considering higher-order responsibility.
Two different forms of responsibility, counterfactual and seeing-to-it, have been extensively discussed in the philosophy and AI in the context of a single agent or multiple agents acting simultaneously. Although the generalisation of counterfactual responsibility to a setting where multiple agents act in some order is relatively straightforward, the same cannot be said about seeing-to-it responsibility. Two versions of seeing-to-it modality applicable to such settings have been proposed in the literature. Neither of them perfectly captures the intuition of responsibility. This paper proposes a definition of seeing-to-it responsibility for such settings that amalgamate the two modalities. This paper shows that the newly proposed notion of responsibility and counterfactual responsibility are not definable through each other and studies the responsibility gap for these two forms of responsibility. It shows that although these two forms of responsibility are not enough to ascribe responsibility in each possible situation, this gap does not exist if higher-order responsibility is taken into account.