ChatGPT, a hype that teaches: generative AI and the collective accumulation of technology literacy on social media

The 2022 launch of ChatGPT triggered a wave of social media discussion about generative AI. Drawing on 215,279 posts from five months after this launch, we extend prior accounts of hype as vehicle for strategic resource mobilization, to develop a theory of hype as collective sensemaking and knowledge creation. Existing models rest on a fixed relational architecture: knowledgeable actors leverage information asymmetries to project visions, directing expectations top-down. ChatGPT collapsed this architecture by making generative AI accessible, democratizing technology evaluation. Meanwhile, social media emerged as the arena where relational publics could form, driving vision formation and contestation around generative AI from bottom-up, reshaping how a nascent technology acquired meaning. Combining grounded theory with computational text analysis, we develop a process model on relational publics’ role in hypes. We explain how collective knowledge accumulates through traceable conceptual prototyping: testing a technology's capability and proposing meaning simultaneously. Also, that when this knowledge crosses a technology literacy threshold, hype transitions from an expanding to a containing phase. The model advances our understanding of how novel technologies diffuse when crowds simultaneously co-construct visionary imaginaries and practical knowledge, and of the role social media plays in this process.