“What People Are Saying” About Extreme Weather: A Comparison of Public Commentary and AI-Generated Summaries

Background: Weather risk communication requires balancing technical data with personal and community contexts. Extreme weather messaging from the US National Weather Service (NWS) via social media inspires both public responses and artificial intelligence-generated summary content. Literature review: While artificial intelligence (AI) is used to process data from meteorological models, there remains a need to assess how AI mediates weather communication. Research in technical communication and meteorology agree that assessing AI communication in context is key. Research questions: 1. How do public responses to NWS forecasts vary across different communities and types of extreme weather? 2. When compared to human responses, what do AI-generated summaries include and exclude to assist with risk decision-making? 3. What are the implications for how technical and professional communication (TPC) should define its relationship with AI? Research methodology: I combine corpus analysis with open and axial coding to compare how human responses differ across weather types and how they compare to AI-generated summaries. Results: Public commenters respond to weather risks along a continuum of active, reactive, and reflective modes, depending on the nature of the post, timeline, and perceived level of risk. Meta-AI summaries share general information, but do not reflect the texture of public responses that bolster building community resilience. However, NWS message type and language choices exert a strong influence guiding public conversation. Conclusion: In the future, TPC scholars and those working in science communication contexts should focus on concrete, community-oriented actions and human connection, relying on AI-generated summaries only for generalized communication contributions.