American and Chinese subjects’ explicit and implicit perceptions of AI-generated content: A mixed-methods approach

As the prevalence of AI-generated content increases, examining viewers’ perceptions of the content is crucial to understand the human-machine relationship and further facilitate efficient human-machine collaboration. Prior literature has accumulated mixed findings regarding subjects’ attitudes toward and perceptions of news and tweets written by natural language generation (NLG) algorithms. To resolve this inconsistency, this study investigated the explicit and implicit perceptions of AI-generated poetry and painting held by subjects from two societies. An online experiment and subsequent content analysis were conducted to examine the subjects’ explicit and implicit perceptions of AI-generated content in the U.S. and China. As the U.S. and China fiercely compete to lead the development of AI technology, their citizens exhibit divergent attitudes toward AI’s performance in NLG and the creation of art. The U.S. subjects were more critical of the AI- than the human-generated content, both explicitly and implicitly. Although the Chinese subjects were overtly positive about the AI-generated content, they appreciated less this content than the human-authored content. The findings enrich our understanding in the domain of AI generation. Theoretical and practical implications are discussed.