摘要
用户生成内容(UserGeneratedContent,UGC)对于理解不同用户群体对新兴技术的态度具有关键作用。通过跨平台内容分析,可全面评估公众对ChatGPT的认知。基于6个社交媒体平台的UGC,通过向量融合构建LDA-SBERT模型,挖掘UGC中的隐含主题特征,并进一步利用中文情感分析模型BERT-wwm分析不同主题、不同平台的公众情感态度。研究表明,ChatGPT相关的UGC可归为经济与产业影响、技术应用与体验、社会关系与影响、安全与伦理风险四大维度14个类别;跨平台比较表明,平台间相关话题讨论既有共性也存在特性,展现了不同的平台特色;情感分析结果显示,由于存在认识视角和未来预期的差异,公众对ChatGPT的态度呈现出明显的两极分化。研究不仅有助于技术开发者和政策制定者更好地理解公众对AI技术的态度,也为社交媒体平台的内容管理和舆情引导提供理论支撑。
User generated content(UGC)plays a key role in understanding the attitudes of different user groups towards emerging technologies.We can comprehensively evaluate the public’s attitudes towards ChatGPT through cross-platform content analysis.Based on the UGC of six social media platforms,the LDA-SBERT model is constructed through vector fusion to deeply mine the implicit topic features in UGC.The Chinese sentiment analysis model BERT-wwm is further used to analyze the public’s emotional attitudes under different topics and platforms.The study shows that ChatGPT-related UGC can be classified into 14 categories in four dimensions:economic and industrial impact,technical application and experience,social relations and impact,and security and ethical risks.Cross-platform comparisons show that the discussion of related topics between platforms has both commonalities and characteristics,showing different platform characteristics.The sentiment analysis results show that due to differences in cognitive perspectives and future expectations,the public’s attitude towards ChatGPT is obviously polarized.This study not only helps technology developers and policymakers better understand public attitudes toward AI technology,but also provides theoretical support for content management and public opinion guidance on social media platforms.
作者
黄本涛
李璨
吴红
郭姗姗
HUANG BenTao;LI Can;WU Hong;GUO ShanShan(School of Medicine and Health Management,Tongji Medical College of Huazhong University of Science and Technology,Wuhan 430030,P.R.China)
出处
《数字图书馆论坛》
2025年第2期35-46,共12页
Digital Library Forum
基金
国家自然科学基金面上项目“在线医疗团队协作模式与绩效提升策略研究”(编号:72371111)
国家自然科学基金青年科学基金项目“基于多渠道融合的医疗服务模式及其影响机制研究”(编号:72001087)资助。
