Artificial intelligence-enabled diagnostics promise to transform hearing healthcare, yet real-world adoption remains limited. This study identifies and prioritizes barriers to AI integration in clinical audiology through a three-phase mixed-methods approach. Phase I reviewed literature, surfacing 20 obstacles across workflow, infrastructure, culture, and ethics. Phase II involved expert interviews, refining these into nine context-specific barriers. In Phase III, a fuzzy-DEMATEL survey and thematic coding revealed a causal hierarchy: algorithmic inaccuracy, privacy concerns, and lack of training erode clinician trust and widen the knowledge gap. Cost, integration issues, and resource limitations add systemic stress, while ethical concerns emerge downstream. Cluster analysis groups the barriers into three blocs: Clinical Workflow, Governance and Trust, and Tech Infrastructure. This is the first study to apply fuzzy-DEMATEL to AI barriers in audiology, producing a causal map and cluster framework that offer both theoretical insights and practical guidance for adoption strategies.