Evaluating the Effectiveness of Artificial Intelligence Language Models in Hearing Loss Assessment: Comparative Study of ChatGPT, Gemini, and Perplexity

Aim: Hearing loss constitutes a considerable health concern that adversely impacts individuals' quality of life. This study aims to compare the efficacy of artificial intelligence (AI) based methods in evaluating hearing loss.Material and Method: The study assessed the performance of the ChatGPT, Gemini, and Perplexity programs based on quality (GQS), accuracy (Likert), and readability (GFI). The data were analyzed utilizing nonparametric tests, and the intergroup differences were assessed with the Kruskal-Wallis H test. Groups exhibiting substantial differences were analyzed using the Bonferroni-adjusted Post-Hoc test.Results: The investigation indicates that ChatGPT outperforms other tools in quality (p=0.018). No substantial difference was seen between the groups regarding accuracy (p=0.072) and readability (p>0.05). The GQS score for ChatGPT was determined to be 4.93, for Gemini it was 4.71, and for Perplexity it was 4.43.Conclusion: ChatGPT has exhibited enhanced performance regarding quality in the assessment of hearing loss. Nonetheless, comparable outcomes regarding accuracy and readability indicate that alternative approaches may also prove effective in some applications. These findings endorse the efficacy of AI-based technologies in particular health concerns, such as the evaluation of hearing loss. In the future, it is advisable to evaluate these technologies using larger samples and across various health conditions.