Comparative Analysis of the Accuracy and Readability of the Answers Given by ChatGPT, DeepSeek and Gemini Artificial Intelligences about Penile Body Dysmorphic Disorder

INTRODUCTION: This study contrasts the validity and legibility of responses of three computer algorithms—ChatGPT, DeepSeek, and Gemini—to Penile Body Dysmorphic Disorder (BDD), where individuals are concerned with penis size. As individuals more and more rely on AI to seek health knowledge, this research contrasts the validity and readability of AI-generated responses to common penile size and function questions.

METHODS: Fourteen frequently asked questions about penis size and function were collected from social media and posed to ChatGPT, DeepSeek, and Gemini. Responses were evaluated with the Discern index to evaluate reliability and the Gunning Fog index to measure readability. For statistical analysis one-way ANOVA and one-way ANOVA and Post hoc tests were used.
RESULTS: There was no significant difference between three algorithms (p=0.063), and the distinct scores for ChatGPT, DeepSeek, and Gemini were 3.29±1.07, 3.86±0.66, and 3.43±0.85. In comparison to DeepSeek (19.81±1.88) and Gemini (20.07±2.37), ChatGPT produced a significantly lower Gunning Fog score of 18.34±1.72 (p=0.012), suggesting that ChatGPT responses were easier to read.
DISCUSSION AND CONCLUSION: The study found all three AI models provided accurate and informative content about Penile BDD, where DeepSeek provided the most precise and ChatGPT the easiest to read. However, AI should never replace expert medical recommendations. Continuous improvement and ethical reasoning are required to ensure safe application of AI for health information.

Keywords: Penis, BDD, ChatGPT