How online communities of practice (CoPs) around artificial intelligence (AI) construct, construe, and shape sociotechnical imaginaries is relatively unexplored, as most studies focus on elite discourse in politics, technology, and the media. More research is needed into specific communities and their views on AI, especially those who use it daily for personal interests. This can reveal to what extent dominant AI imaginaries in public discourse are adopted or challenged in CoPS. The present study critically explores an online community engaged in AI-generated livestreams. CoP members collaboratively explore AI’s potential for audio-visual creativity, deeply immersing themselves in AI mediatisation, where they share, discuss, and generate content with AI. Specifically, the study examines the Discord community Nothing, Forever. The empirical analysis combines computational and qualitative methods for a frame analysis of a large corpus of community conversations (N = 7,934) to uncover how participants discuss the technology. The results show that the community’s AI imaginary is dominated by four themes: AI as a Tool for Creative Practices, Human−AI Interaction, Technical Foundations and Ethics, Labor and Social Impact. While community members primarily focus on practical aspects of AI for content creation, context-specific discussions link to themes of technological potential and promises. However, ethical considerations are often left out and questions of AI governance are rarely raised.
