Examining scholarly communication on X (Twitter): insights from participants tweeting COVID-19 and ChatGPT publications

This study explores the dynamics of online scholarly communication through the lens of diffusion of innovation theory, examining participant reactions and interactions surrounding publications on two trending topics, COVID-19 and ChatGPT, on X (formerly Twitter). Employing a customized automated user classifier, we analyze behaviors across diverse user groups using a dataset comprising 415,492 X users. Our findings indicate that scholarly communication on X is heavily shaped by the broader social context. Discussions about COVID-19 publications, driven by the urgency of a public health crisis, attracted a wider range of participants. The prevalence of @mentions and replies in relevant discussions underscores community-driven engagement during the pandemic. In contrast, ChatGPT-related publications, focused on artificial intelligence and machine learning, primarily engaged academic and professional communities. Discussions surrounding scholarly works on X may also be influenced by the platform's algorithms, which prioritize content that prompts immediate and rapid reactions. Our study is among the first to analyze temporal patterns of user reactions, identifying a peak in discussions shortly after publication releases, followed by a rapid decline. While participants responded more quickly to COVID-19 publications, these discussions exhibited a shorter lifespan compared to those related to ChatGPT. In general, user interactions within X-based scholarly communication are initiated by conversations among academic publishers, researchers, and health science practitioners, extending to a broader audience during peak periods. Although discussions on X may not sustain prolonged engagement due to their relatively short span, it is promising to observe sustained connections between academia and professional communities in later stages, potentially fostering a translational impact of research.