Explainable Artificial Intelligence (XAI) provides human understandable explanations into how AI systems make decisions in order to increase transparency. We explore how transparency levels in XAI influence perceptions of fairness, trust and understanding, as well as attitudes towards AI use. The transparency levels – no explanation, opaque, simple and detailed - were varied in two contexts - treatment prioritization and recidivism forecasting. In eight experimental groups, 573 participants judged these explanations. As predicted opaque explanations decreased trust and understanding, but surprisingly simple explanations that were more limited in the information they provided had stronger effects on trust and understanding than detailed explanations. Transparency levels did not have an impact on perceptions of fairness and attitudes towards AI, but context did, with the recidivism AI being perceived as less fair. The findings are discussed in relation to information overload and task subjectivity vs objectivity.

We introduce a novel methodology to scrutinize blame attributions in 'Tweets', focusing on Artificial Intelligence (AI) incidents - a contemporary issue that provokes regular discourse. The method identifies the agents that get blamed and the factors that are associated with blame attributions. The proposed methodology replicates and contextualizes findings from experimental settings, revealing AI entities are often held accountable for adverse outcomes, while human agents are judged based on intentions. It also identifies unexplored factors, such as blaming data for perceived biases or AI for replacing humans. This method offers a robust tool for mitigating measurement bias in specific fields, enabling the continual rejuvenation of theoretical frameworks with emerging variables.

This study investigates how people assign blame to autonomous vehicles (AVs) when involved in an accident. Our experiment (N = 2647) revealed that people placed more blame on AVs than on human drivers when accident details were unspecified. To examine whether people assess major classes of blame-relevant information differently for AVs and humans, we developed a causal model and introduced a novel concept of prevention effort, which emerged as a crucial factor for blame judgement alongside intentionality. Finally, we addressed the “many hands” problem by exploring how people assign blame to entities associated with AVs and human drivers, such as the car company or an accident victim. Our findings showed that people assigned high blame to these entities in scenarios involving AVs, but not with human drivers. This necessitates adapting a model of blame for AVs to include other agents and thus allow for blame allocation “outside” of autonomous vehicles.