The hype effect: How exaggerated AI science news shapes perceptions of scientists and journalists

This study investigates how varying levels of hype—defined as a rhetorical framing strategy involving exaggeration or amplification—and the valence of portrayed outcomes in AI news coverage influence audience perceptions of the credibility of scientists, journalists, and science news sources. A total of 622 U.S.-based participants were randomly assigned to view one of ten video segments, which varied in topic (self-driving cars or finance), hype intensity (neutral, low, high), outcome valence (positive or negative), and source (neutral, international broadcaster, or scientific institution). Post-exposure measures assessed emotional responses (e.g., anxiety), message evaluation, and credibility perceptions. Findings indicate that hype intensity and valence had limited direct effects on the perceived credibility of scientists, suggesting that public trust in science may depend more on emotional responses and source characteristics than on sensationalist framing alone. Conversely, journalist and source credibility were more sensitive to individual attitudes to AI, emotional responses, and source characteristics. Notably, heightened anxiety predicted lower credibility ratings across all source types, underscoring the mediating role of affect. Gender differences also emerged, with women rating sources as more trustworthy than men. These results highlight the complex interplay between message framing, emotional responses, and pre-existing attitudes in shaping credibility assessments, emphasizing the need for nuanced science communication strategies that consider audience diversity and emotional engagement.