Generative Artificial Intelligence (GenAI) is increasingly integrated into scientific work (e.g., Kwon, 2025). In the research on public engagement with science, GenAI could be used to assist scientists in helping the public understand science in meaningful ways. For example, Hendriks et al. (2025) suggested that good working habits for using GenAI in science communication consider ethics and effectiveness. Feldman et al., (2025) also cautioned that GenAI use in science communication requires serious consideration of ethical issues surrounding GenAI technology, such as infrastructure, governance, and data. We address one of the research agendas proposed by Kessler et al. (2025) about GenAI and science communication research: communication with AI.
Focusing on climate science communication on social media, where interactive affordances enable direct dialogue between scientists and the public (Hara & Chae, 2025), we investigated scientists' perspectives on using GenAI to communicate climate science in mediated environments. The presentation will discuss a survey conducted among climate scientists regarding their use of GenAI for public engagement with science when creating social media content. The survey asks climate scientists’ experiences with GenAI-assisted science communication, their perceptions of its benefits and risks, and their practices toward quality control of GenAI output, which includes the issue of how to maintain public trust and credibility during a time when science’s legitimacy is being challenged (Gauchat, 2023). By focusing on scientists' own accounts of how they assess, revise, and employ GenAI-generated communication, this study explores the roles of GenAI in public engagement with science in mediated environments.