Advances and Challenges in Artificial Intelligence and Machine Learning for Multilingual Speech Disorder Diagnosis and Therapy: An Integrative Review

This study examines the role of Artificial Intelligence (AI) and Machine Learning (ML) in the diagnosis, assessment, and management of speech disorders, focusing on enhancing clinical practices. A comprehensive review of existing studies was conducted, highlighting the application of AI technologies such as Automated Speech Recognition (ASR), Natural Language Processing (NLP), and deep learning models, including Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs). The findings reveal that AI and ML techniques have significantly improved diagnostic precision, therapeutic interventions, and clinical efficiency, especially in underserved populations. However, challenges related to limited multilingual datasets, model generalizability, and the interpretability of deep learning models were identified. Despite the promising advancements, challenges such as data privacy, model bias, and ethical concerns need to be addressed for broader clinical integration. This study contributes by synthesizing current AI and ML applications in speech disorder management, identifying key challenges, and proposing future directions, including the development of multilingual datasets, Explainable AI (XAI), and the integration of multimodal data to further enhance diagnostic and therapeutic outcomes.