Readability Analysis of AI-Generated Medical Content: Comparison of ChatGPT and Gemini in Diverticulitis
Acta Medica Young Doctors · 12.04.2026 · 2 (2)
DOI: 10.66288/actamedi.2026.80 ↗Kurz gefasst
Fachartikel in der Zeitschrift Acta Medica Young Doctors, online erschienen am 12.04.2026. Verfasst von Resmiye Nur Okudan. Die Arbeit nennt 28 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Background
Diverticulitis is a common gastrointestinal condition for which patients increasingly seek information. Large language models (LLMs), such as ChatGPT and Gemini, are widely used to generate medical texts. However, the readability of these outputs remains a critical concern for effective patient education.
Objective
This study aimed to evaluate and compare the readability of responses generated by ChatGPT and Gemini for commonly asked questions about diverticulitis.
Methods
A set of standardized diverticulitis-related questions reflecting common patient concerns was developed and submitted to both models for evaluation. The generated responses were analyzed using multiple validated readability indices, including the Automated Readability Index (ARI), Flesch Reading Ease (FRE), Gunning Fog Index (GFI), Flesch-Kincaid Grade Level (FKGL), Coleman–Liau Index (CLI), SMOG Index, and Linsear Write Formula. An aggregate measure, the Average Reading Level Consensus (ARLC), was also calculated. Statistical comparisons were performed, with p < 0.05 considered significant.
Results
Both ChatGPT and Gemini produced responses with readability levels exceeding the recommended standards for patient education. The ARLC was higher for ChatGPT than Gemini (11.45 ± 3.1 vs. 10.5 ± 1.6; p = 0.322), although the difference was not statistically significant. Gemini demonstrated significantly better readability based on the FRE score (32.4 ± 10.2 vs. 28.22 ± 10.4; p = 0.027). No significant differences were observed in ARI, GFI, FKGL, CLI, SMOG, or the original Linsear Write Formula (p > 0.05). However, the Linsear Write Grade Level was significantly higher for ChatGPT (20.01 ± 8.9 vs. 7.8 ± 2.1, p < 0.001).
Conclusions
Both LLMs generated diverticulitis-related content at readability levels above the recommended thresholds, potentially limiting patient comprehension. Gemini generally produced slightly more accessible text than ChatGPT. These findings highlight the need to optimize AI-generated medical content to better align with health literacy standards and improve patient understanding.
Zitieren
APA
Okudan, R. N. (2026). Readability Analysis of AI-Generated Medical Content: Comparison of ChatGPT and Gemini in Diverticulitis. Acta Medica Young Doctors, 2(2). https://doi.org/10.66288/actamedi.2026.80
Vancouver
Okudan RN. Readability Analysis of AI-Generated Medical Content: Comparison of ChatGPT and Gemini in Diverticulitis. Acta Medica Young Doctors. 2026;2(2). doi:10.66288/actamedi.2026.80
BibTeX
@article{Okudan2026,
author = {Okudan, Resmiye Nur},
title = {Readability Analysis of AI-Generated Medical Content: Comparison of ChatGPT and Gemini in Diverticulitis},
journal = {Acta Medica Young Doctors},
year = {2026},
volume = {2},
number = {2},
doi = {10.66288/actamedi.2026.80}
}
EUVIMED