<p>With the era of digital revolution, artificial intelligence has become a critical point for human life. Large language models (LLMs), which is widely used especially in the field of health and medicine, is used as a source of information by patients and their families as well as diagnosis and treatment. The aim of this study is to evaluate the accuracy, reliability, and usability of information provided by ChatGPT when asked questions frequently posed by patients and their families regarding scoliosis and its treatment, considering machine learning and language barriers. In the study, frequently asked questions about scoliosis and its treatment were discussed by two expert physiotherapists, resulting in a final list of 20 questions. Each question was asked in two different ChatGPT sessions and languages. The responses provided were evaluated on a Likert scale in terms of validity, safety, and usefulness. Seventy-five percent of the responses generated by ChatGPT received full scores in both the native language and in English. Responses varied across different session or language conditions, though no significant differences were detected. Validity, safety, and usefulness scores were similar across all categories (Md = 4.0, <i>p</i> &gt; 0.05). The differences in scores for responses to conservative and surgical treatments are remarkable. This study has demonstrated that ChatGPT can be a useful digital resource for accessing information. However, it may be more appropriate to use ChatGPT following expert consultation when seeking general information about a medical condition.</p>