Communicative AI in the scientific public sphere: An analysis of Twitter discourse on generative AI tools

Abstract

Drawing on the concept of the scientific public sphere, this study examines the public sense-making of communicative AI (e.g., generative AI) on social media. Advancing a framework encompassing cognitive (technology vs. use) and affective (positive vs. negative) dimensions of the public discourse on communicative AI, we analyzed global Twitter (now X) conversations about generative AI tools. Findings showed that the text generator (ChatGPT) discussions centered more on the technology-centered themes, whereas the image generator discussions emphasized their uses. ChatGPT received mixed sentiments in technology-related discussions, while there was more positive sentiment about the its uses. Theoretical and practical implications are discussed.