Artificial Intelligence in Climate Change Journalism: Assessing Perceived News Credibility, Readability, and Influence on Climate Change Mitigation Intention

As artificial intelligence-generated (AI-generated) journalism rises, its effectiveness in engaging the public to mitigate climate change remains uncertain. This study integrates the MAIN model and expectation-confirmation theory to examine how AI-generated versus human-generated climate change news influences audience news evaluation and behavioral intentions. We conducted a 2 (human-generated news vs. AI-generated news) × 2 (authorship disclosed vs. authorship undisclosed) × 2 (narrative vs. non-narrative) between factorial experiment (N = 441) to test the effects of news source, authorship disclosure, and narrative style on readers’ perceived news credibility, readability, and behavioral intentions for climate change mitigation. The t-tests revealed that human-generated climate change news elicited positive disconfirmation, while AI-generated news resulted in negative disconfirmation. A three-way MANCOVA revealed that narrative news significantly promoted climate change mitigation intentions. Further moderation analyses indicated that while AI-generated news did not differ significantly from human-generated news in terms of persuasiveness, improving the algorithmic narrative structure of AI-generated content could enhance positive disconfirmation of readability, thereby increasing the likelihood of engagement in mitigation behaviors. These findings provide insights into optimizing AI-generated journalism to enhance climate change communication and public engagement.