'Put the Car on the Stand': SMT-based Oracles for Investigating Decisions

Principled accountability in the aftermath of harms is essential to the trustworthy design and governance of algorithmic decision making. Legal philosophy offers a paramount method for assessing culpability: putting the agent 'on the stand' to subject their actions and intentions to cross-examination. We show that under minimal assumptions automated reasoning can rigorously interrogate algorithmic behaviors as in the adversarial process of legal fact finding. We model accountability processes, such as trials or review boards, as Counterfactual-Guided Logic Exploration and Abstraction Refinement (CLEAR) loops. We use an SMT-based oracle to discharge queries about agent behavior in factual and counterfactual scenarios, as adaptively formulated by a human investigator. For a decision algorithm $\mathcal{A}$, we use symbolic execution to represent its logic as a statement $\Pi$ in the decidable theory $\texttt{QF_FPBV}$. We implement our framework in a tool called $\textsf{soid}$ with an accompanying GUI, and demonstrate its utility on an illustrative car crash scenario.