Towards Responsible AI: Developing Explanations to Increase Human-AI Collaboration

Most current XAI models are primarily designed to verify input-output relationships of AI models, without considering context. This objective may not always align with the goals of Human-AI collaboration, which aim to enhance team performance and establish appropriate levels of trust. Developing XAI models that can promote justified trust is therefore still a challenge in the AI field, but it is a crucial step towards responsible AI. The focus of this research is to develop an XAI model optimized for human-AI collaboration, with a specific goal of generating explanations that improve understanding of the AI system's limitations and increase warranted trust in it. To achieve this goal, a user experiment was conducted to analyze the effects of including explanations in the decision-making process on AI trust.