生成式AI交互情境下用户隐私风险自适应模型构建OACHSSCD

[Purpose]In the context of generative artificial intelligence(Gen AI)interactions,users'privacy information disclosure behav-iors exhibit highly complex,contextualized,and dynamic characteristics.A deep exploration of the entire process of privacy disclosure in users'sustained interactions with Gen AI can reveal the intricate psychological and behavioral mechanisms underlying privacy decision-making in human-AI interaction,thereby providing critical theoretical support for building trustworthy AI systems.[Method]Drawing on privacy calculus theory,protection motivation theory,and the stimulus-organism-response model as theoretical frameworks,data were collected through semi-structured interviews.Thematic analysis was employed,combining manual coding with GPT-assisted coding,to i-dentify core themes of users'adaptive processes to privacy risks in Gen AI contexts.[Result/Conclusion]A dynamic integrative model of user privacy disclosure in Gen AI interactions is proposed,comprising three core components:one is the triggering factors,including social information,privacy policy,empirical threat,and situational demand;the second is the decision-making process,including privacy inva-sion risks perception,functional benefit perception,and coping appraisal;the third is the behavioral response,including adaptive and non-adaptive coping strategies.The findings advance the understanding of privacy disclosure behaviors in human-AI interactions and offer im-plications for designing more privacy-aware Gen AI systems.