Public emotional dynamics toward AIGC content generation across social media platform
arxiv(2023)
摘要
Given the widespread popularity of interactive AI models like ChatGPT, public
opinion on emerging artificial intelligence generated content(AIGC) has been
extensively debated. Pessimists believe that AIGC will replace humans in the
future, and optimists think that it will further liberate productivity. Public
emotions play a crucial role on social media platforms. They can provide
valuable insights into the public's opinions, attitudes, and behaviors. There
is a lack of research on the analysis of social group emotions triggered by
AIGC content, and even more on the cross-platform differences of group
emotions. This study fills the research gap by connecting the theory of group
dynamics with emotions in social media. Specifically, we develop a scientific
group emotion calculation and visualization system based on chains of
communication. The system is capable of crawling data in real time and
presenting the current state of group emotions in a fine-grained manner. We
then analyze which group dynamic factors drive different public emotions
towards nine AIGC products on the three most popular social media platforms in
China. Finally, we obtain four main findings. First, Douyin is the only
platform with negative group emotion on emerging AI technologies. Second, Weibo
users prefer extreme emotions more than others. Third, the group emotion varies
by education and age. It is negatively correlated with senior high school or
lower and 25 or younger, and positively correlated with bachelor's degree or
higher and 26-35. Fourth, the group emotion polarization increases with more
posts without comments and celebrity publishers. By analyzing the key dynamic
factors of group emotions to AIGC on various social media platforms, we can
improve our products and services, develop more effective marketing strategies,
and create more accurate and effective AI models to solve complex problems.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要