Global utilization of artificial intelligence in the diagnosis and management of voice disorders over the past five years

Abstract Objective This review evaluates the worldwide use of artificial intelligence (AI) for the diagnosis and treatment of voice disorders. Methods An electronic search was completed in Embase, Pubmed, Ovid MEDLINE, Scopus, Google Scholar, and Web of Science. Studies in English from 2019 to 2024 evaluating the use of AI in detection and management of voice disorders were included. Preferred Reporting Items for Systematic Reviews and Meta‐Analyses guidelines were followed. Results Eighty‐one studies were recognized. Thirty‐three studies were chosen and screened for quality assessment. Of these, 16 studies used AI to determine normal versus pathological voice. The convolutional neural network (CNN) was the most employed algorithm among all machine learning algorithms. Conclusion This review revealed significant interest worldwide in utilizing AI in detection of voice disorders. Gaps included the use of limited, inconsistent data sets, lack of validation, and emphasis on detection rather than treatment of the voice disorder. These are areas of opportunity for AI techniques to improved diagnostic accuracy.