Abstract: 【Purpose/significance】The technical dividend of generative artificial intelligence gives criminals insight into the feasibility and wide ′money prospect′ of fraudulent means using GAI to produce information fog, and it is particularly critical to clarify the overall process of information fog occurrence in the fraud situation to effectively block GAI fraud crimes【. Method/process】This paper aims to conduct an exploratory research on the formation mechanism and audience response of information fog in GAI fraud scenario. We se⁃ lected GAI fraud reports and interview texts as data sources, and used systematic thematic analysis method to data analyze.【Result/ conclusion】(1)for the formation mechanism of information fog, four influencing factors are information environment, information tech⁃ nology, information and information subject. Among them, information environment presents the characteristics of information overload and social network, information technology presents deepfakes, computational propaganda and anthropomorphic, information presents ′credible′ disinformation, and information subject has cognitive bias, digital literacy and awareness of prevention. Guided by informa⁃ tion ecology theory, we further explains the mechanism of information fog formation. (2)for the audience response of information fog, we obtain three dimensions of cognition, behavior and emotion. Among them, cognition leads to discrimination and prejudice, emotion ap⁃ pears social trust loss and anxiety, and behavior has privacy disclosure and be deceived. Guided by cognitive behavior theory, we ex⁃ plains the intrinsic process of audience response【. Innovation/limitation】This study enriches the research of information fog to a certain extent, and provides reference for GAI fraud management. However, the richness of GAI fraud samples and the applicability of theo⁃ retical models need to be further developed and verified.