ABSTRACT
Accurate diagnosis of mental illness is challenging due to the absence of apparent physical symptoms, overlapping disorders, and the stigma surrounding mental health. Traditional diagnostic methods are often inaccessible, making artificial intelligence (AI) a promising tool to enhance diagnostics. However, building trust in AI requires developing explainable AI (XAI) to help physicians understand AI decisions and improve public acceptance. This chapter examines obstacles, biases, and the adoption of XAI in psychiatry through a systematic literature review. We analyze AI techniques, algorithms, data collection methods, and explainability approaches. Our findings highlight the potential of XAI in mental healthcare while emphasizing the need for solutions to enhance transparency, interpretability, and trust in AI-driven diagnostics.
