WHEN PRIVACY CONCERNS DON’T STOP DATA SHARING: EXPLAINING THE DATA PRIVACY PARADOX IN AI-POWERED PERSONALIZED MARKETING

Abstract

The integration of artificial intelligence (AI) into digital marketing has increased a paradox of consumer behaviour that has been present for years: consumers often voice significant privacy concerns but are still willing to provide platforms, brands, recommendation systems, conversational agents and personalised advertising technologies with personal information. The data privacy paradox is especially complicated in personalized marketing with AI algorithms, as AI can turn your usual behaviors into intricate forecasts about consumer preferences, intentions, identities and future behavior. The focus of the present study is to investigate why privacy-related issues do not necessarily deter consumers from disclosing their personal data in AI-assisted marketing interactions. The study employs a conceptual research design that involves an integrative review of the foundational and contemporary research literature, synthesizing privacy calculus theory, contextual integrity, trust theory, technology acceptance approaches, uses and gratifications, personalization research, behavioral economics and the recent human–AI interaction literature. The analysis reveals five interrelated mechanisms that contribute to this paradox: perceived benefits of personalization, trust and perceived control, contextualization of data disclosure, asymmetry between immediate and future perceived threats, and increasing opacity and inferential power of the AI systems. The results show that the “privacy paradox” is not just irrational inconsistency in attitudes and behaviour. Instead, data sharing is a process of negotiation, with consumers balancing privacy concerns with the benefits of privacy, utility, utility efficiency, utility gratification, social involvement, and feelings of control. The study also suggests that AI has also shifted the focus from privacy paradox (disclosure of information) to algorithmic inference (inference of information). The integrative framework that is developed in the discussion links privacy concern indirectly to disclosure via perceived risk, personalization value, trust, control, contextual appropriateness and privacy-risk salience. The authors conclude that responsible AI-marketing needs to go beyond "notice-and-consent" to focus on personalisation that is transparent, contextual, proportionate, explainable and respects privacy.