Can ChatGPT be a dependable resource for parents looking for information on specific learning disorder?
TL;DR
The findings indicate that ChatGPT generally provides accurate and comprehensive information on SLDs, however, parents’ responses often included technical terminology and excessive detail, reducing clarity, and reinforcing ChatGPT’s role as a supplementary tool rather than an independent medical information source for non-expert users.
Abstract
ChatGPT is an innovative artificial intelligence (AI) tool that is increasingly applied in healthcare and is widely used by professionals and the public for medical information. This study evaluated ChatGPT’s capacity to provide reliable, evidence-based information on specific learning disorders (SLDs) to parents. A set of questions about specific learning disorders, which are based on questions frequently asked by patients and their families in psychiatric interviews and supported by relevant literature, was prepared by two experienced child and adolescent psychiatry specialist. Two mental health professionals independently evaluated the responses. The findings indicate that ChatGPT generally provides accurate and comprehensive information on SLDs. However, parents’ responses often included technical terminology and excessive detail, reducing clarity. Although ChatGPT has the potential to deliver accurate and detailed information, it may cause misunderstandings or undue concerns for nonspecialists. This study evaluates the accuracy, clarity, and inclusivity of ChatGPT’s responses to questions posed by different user profiles—a parent/caregiver and an experienced expert. Results indicated that ChatGPT demonstrated consistently high accuracy, with median accuracy scores of 5 and mean scores of 4.8 and 4.7 for responses given to parents/caregivers and experts, respectively. No significant difference was found between the two groups. Similarly, inclusivity scores were high across both groups, with median values of 5 and mean scores of 4.82 and 4.87, respectively, showing no significant difference. However, clarity and comprehensibility varied by user group. These findings emphasize the need for refinements to enhance ChatGPT’s adaptability to different levels of user expertise. Ensuring that responses are tailored appropriately for lay audiences can improve clarity while maintaining accuracy. Future developments should focus on optimizing language simplification and contextual adaptation, reinforcing ChatGPT’s role as a supplementary tool rather than an independent medical information source for non-expert users.