Abstract
Introduction Artificial intelligence tools such as ChatGPT and Google Gemini are increasingly used for generating patient education materials. This study aimed to compare the readability, reliability, and originality of artificial intelligence (AI)-generated patient education content on surgical management options for breast cancer. Methods A cross-sectional study was conducted from January 15 to 21, 2025. Standardized prompts requesting patient education brochures for common breast cancer surgeries were given to ChatGPT-4o and Gemini 2.0 Flash. All responses were generated in English. Readability was assessed using the validated Flesch-Kincaid Readability Calculator, originality using a similarity checker, and reliability using the Modified DISCERN tool. Statistical comparisons were performed using unpaired t-tests with significance set at p<0.05. Results There were no significant differences between the two AI tools in syllables per word, ease score, similarity, or reliability. ChatGPT produced significantly higher word counts (p=0.018) and sentence counts (p=0.001). Gemini generated longer sentences and a higher grade-level score (p=0.002), indicating relatively more complex text. Conclusions Both AI tools produced patient education materials with broadly comparable readability, reliability, and originality. No correlation was observed between readability ease and reliability scores, suggesting consistent performance across both platforms.