IXAII: An Interactive Explainable Artificial Intelligence Interface for Decision Support Systems

One-size-fits-all explanations don't work: The paper found that static AI explanations fail to meet diverse user needs, demonstrating that different stakeholders (such as domain experts, managers, and general users) require tailored information to understand AI decisions effectively.

Multiple explanation methods provide better understanding: IXAII combines four different explainable AI techniques (LIME, SHAP, Anchors, and DiCE), and the research shows that offering users multiple explanation approaches—rather than just one—increases perceived transparency and helps people grasp how AI systems reach their conclusions.

Users need control over explanations: The system allows users to customize both the content and format of explanations they receive, reflecting a key finding that giving people agency over how AI decisions are explained to them enhances the usefulness of those explanations for decision-making.

Interactive design matters for responsible AI: By moving beyond static explanations to interactive interfaces with visualization options, the research demonstrates that how explanations are presented—not just what information is provided—significantly impacts whether users can actually trust and understand AI systems in practice.

Bridging the gap between theory and practice: The paper addresses a critical gap where explainable AI methods exist in research but aren't effectively implemented in real-world decision support systems, showing that responsible AI requires thoughtful interface design that connects technical methods to actual user needs.