[目的/意义] 探究生成式人工智能的发展给网络信息内容生态机制带来的影响，为优化网络信息内容生态、提升信息利用效能提供作用。[方法/过程] 采用扎根理论研究方法，对生成式人工智能和网络信息内容生态机制相关的文献及政策文件进行开放式编码、主轴编码及选择性编码，构建包含主体机制、参与机制、实现机制和保障机制的生成式人工智能背景下的网络信息内容生态机制。然后，对网络信息内容生态机制内部的相互影响关系进行分析，并解析生成式人工智能对网络信息内容机制的影响路径，在此基础上选择自媒体平台AI应用情况作为案例，验证4条影响路径的合理性。[结果/结论] 研究发现，生成式人工智能对主体机制、参与机制、实现机制、保障机制存在4条影响路径，整体分析逻辑为生成式人工智能—信息内容—信息主体—信息服务—信息环境，在自媒体平台的案例中都能得到有效验证，印证了各路径的合理性与可行性，也充分体现了生成式人工智能对信息内容影响的系统性与动态性。最后，为优化生成式人工智能背景下的网络信息内容生态机制提出一些建议：需规范生成合成内容的传播、严格审查生成模型语料库的选择、强化用户信息判断和信息筛选的能力、明确生成内容涉及的版权问题。

[Purpose/Significance] This paper explores the impact of the development of generative artificial intelligence on the ecological mechanism of network information content, and provides a role in optimizing the ecological mechanism of network information content and improving the efficiency of information utilization. [Method/Process] This paper adopted the grounded theory research method to carry out open coding, axial coding and selective coding on the literature and policy documents related to the ecological mechanism of generative artificial intelligence and network information content, and construct the ecological mechanism of network information content under the background of generative artificial intelligence including subject mechanism, participation mechanism, realization mechanism and guarantee mechanism. Then, the mutual influence relationship within the ecological mechanism of network information content was analyzed, and the influence path of generative artificial intelligence on the mechanism of network information content was analyzed. On this basis, the application of self-media platform AI was selected as a case to verify the rationality of the four influence paths. [Result/Conclusion] The study finds that the overall analytical logic of generative artificial intelligence with respect to the subject mechanism, participation mechanism, realization mechanism, and guarantee mechanism follows the chain: generative artificial intelligence-information content-information subjects-information services-information environment. Moreover, all four influence pathways have been effectively validated through case studies on self-media platforms. This not only confirms the rationality and feasibility of each pathway, but also fully demonstrates the systematic and dynamic nature of generative artificial intelligence's impact on information content. Finally, some suggestions are put forward to optimize the ecological mechanism of network information content under the background of generative artificial intelligence: it is necessary to standardize the dissemination of generated synthetic content, strictly review the selection of generated model corpus, strengthen the ability of user information judgment and information screening, and clarify the copyright issues involved in generated content.