摘要：

【目的】 探讨生成式人工智能(generative artificial intelligence, GenAI)在医学期刊健康科普多模态矩阵中的应用场景和价值。【方法】 采用文献调研法分析GenAI在健康科普文本、图像、音频、视频等模态的应用现状,选取DeepSeek、豆包、即梦AI等GenAI主流模型进行医学科普多模态转换的应用实践,并分析应用成效。【结果】 GenAI在医学期刊健康科普多模态中的应用处于初步阶段但已展现出显著优势。以DeepSeek为代表的大语言模型可优化科普内容;豆包等图像生成模型可高效创作科普插图、科普海报、科普IP形象及表情包等;即梦AI等视频生成模型可创作视频素材;剪映等智能视频工具可大幅提升科普短视频制作效率;蝉镜等数字人技术能为医学专家打造数字分身。【结论】 GenAI可应用于健康科普多模态的诸多场景,显著提高医学期刊的科普内容生产效率,有效提升传播效果,但要确保内容准确性,规避潜在的伦理风险。

关键词: 生成式人工智能, 医学期刊, 健康科普, 多模态, 应用

Abstract:

[Purposes] To explore the application scenarios and value of generative artificial intelligence (GenAI) in the multimodal matrix of health science popularization in medical journals. [Methods] The literature review method was used to analyze the current application status of GenAI in health science popularization text, images, audio, video and other modalities. Mainstream GenAI models such as DeepSeek, DouBao, and Dreamina AI were selected for the application practice of multimodal transformation in medical science popularization, and the application effectiveness was analyzed. [Findings] The application of GenAI in multimodal health science popularization in medical journals is still in its early stages, but it has shown significant application value. Large language models represented by DeepSeek can optimize popular science content; Image generation models such as DouBao can efficiently create popular science illustrations, posters, journal IPs, and emoticons; Dreamina AI and other video generation models can create video materials; Intelligent video tools such as Jianying can significantly improve the efficiency of science popularization short video production; Digital human technologies such as Chanjing can create digital avatars for medical experts. [Conclusions] GenAI can be applied to various scenarios of multimodal health science popularization, significantly improving the production efficiency of science popularization content in medical journals and effectively enhancing the dissemination effect. However, it is necessary to ensure the accuracy of content and avoid potential ethical risks.

Key words: Generative artificial intelligence, Medical journals, Health science popularization, Multimodal, Application