Evaluating the usefulness of ChatGPT and Gemini in generating educational content for patients with vitiligo: a comparative cross-sectional study

Background This study aimed to evaluate the usefulness of ChatGPT and Gemini in generating information suitable for developing educational materials for patients with vitiligo.Methods Information was obtained from ChatGPT and Gemini in response to questions regarding vitiligo derived from the American Academy of Dermatology website. Differences between the groups were analyzed using the Mann-Whitney U test.Results Author 1 assigned a completeness score of 2 to 6 out of 15 responses generated by ChatGPT. All texts received full scores for accuracy, comprehensibility, and usefulness. Six responses from Gemini were scored 2 for completeness, and 1 response received a score of 2 for comprehensibility; all other evaluations were assigned full scores. Author 2 assigned a completeness score of 2 to 3 responses from each model, and 1 Gemini response was rated 2 for comprehensibility. All other assessments received full scores. The mean Flesch Reading Ease scores for ChatGPT and Gemini were 34.92 ± 12.28 and 31.40 ± 13.33, respectively (P = 0.41).Conclusion ChatGPT and Gemini provided largely accurate and useful information on vitiligo; however, most responses were difficult to read. The content should be reviewed and refined by dermatologists to ensure adequate readability before being disseminated to patients.