Comparative analysis of artificial intelligence tools for the dissemination of colorectal cancer screening guidelines: a novel perspective on early screening education

Abstract

This study systematically evaluated the effectiveness of three artificial intelligence (AI) tools-ChatGPT-4o, Claude3.5, and DeepSeek-in disseminating colorectal cancer screening guidelines to non-medical populations. Using uniform instructions aligned with the Chinese society of clinical oncology 2024 standards, the AI-generated content was analyzed for accuracy, clarity, and rigor, supplemented by a cross-evaluation mechanism to quantify performance. Key findings revealed that DeepSeek demonstrated superior regional adaptability and logical rigor, while requiring improvements in threshold accuracy; ChatGPT-4o exhibited outdated starting age criteria and oversimplified high-risk population screening protocols; and Claude3.5 provided a comprehensive framework but lacked critical implementation details. All tools effectively translated complex medical guidelines into accessible language, underscoring AI's potential in public health education. However, outputs necessitated clinical validation and ethical oversight to mitigate data biases. The study emphasizes AI's role as an auxiliary tool for medical knowledge dissemination, advocating for continuous algorithmic optimization, multidisciplinary collaboration, and dynamic regulatory mechanisms to ensure alignment with evolving medical standards while balancing scientific precision and public accessibility.

Keywords: artificial intelligence; chatGPT-4o, claude3.5; colorectal cancer; deepSeek; large language models.