Privacy concerns don’t stop AI use; trust and perceived benefits do

A new study by Seunga Venus Jin reveals that the rapid rise of ChatGPT-powered virtual influencers is disrupting consumer behavior in digital marketplaces, but not without triggering complex tensions around trust, privacy, and social psychology. The research finds that users' willingness to engage with AI-driven recommendations is not determined by technology alone but by a layered interplay of trust, perceived benefits, emotional needs, and concerns over data privacy.

The study, titled "In ChatGPT-Powered Virtual Influencers We (Dis)Trust? The Privacy Paradox and the Double-Edged Sword of Ubiquitous Large Language Model (LLM) Generative AI as a General Purpose Technology (GPT) in a Human-Centered AI Ecosystem," published in Behavioral Sciences, explores how AI equality, trust in ChatGPT, the need to belong, and privacy concerns collectively shape consumers' intention to purchase products promoted by AI-generated influencers.

Trust: The core driver of AI-powered consumer decisions

The study identifies trust in ChatGPT as the single most important factor influencing whether consumers are willing to act on recommendations made by AI-powered virtual influencers. Users who perceive AI systems as reliable, secure, and beneficial are significantly more likely to engage with AI-generated marketing content and make purchasing decisions based on those interactions.

Trust does not emerge in isolation, it is strongly shaped by what the study defines as "AI equality," or the extent to which users feel they can effectively access and benefit from AI technologies in their daily lives. Individuals who perceive themselves as capable of leveraging AI tools for real-world advantages, such as better deals, convenience, or efficiency, show higher levels of trust in ChatGPT systems.

This relationship creates a cascading effect. Higher AI equality leads to greater trust, which in turn increases purchase intention. The study confirms that trust acts as a mediating mechanism, translating technological capability into behavioral outcomes.

The findings also highlight a broader structural issue. The growing divide between those who can effectively use AI and those who cannot is not just a technological gap but a behavioral one. Users on the disadvantaged side of this divide are less likely to trust AI systems and, consequently, less likely to benefit from AI-driven commerce.

This dynamic reinforces concerns that the expansion of AI technologies could deepen existing inequalities, creating a scenario where those already equipped with digital skills gain disproportionate advantages in the marketplace.

Emotional needs and social belonging shape AI influence

The study draws focus to the critical role of human psychology in shaping responses to AI-powered systems. One of the most significant moderating factors identified is the "need to belong," a fundamental human motivation tied to social connection and emotional well-being.

Individuals with a higher need for social belonging are more susceptible to the influence of ChatGPT-powered virtual influencers, particularly in contexts where access to AI resources is limited. In scenarios marked by a strong AI divide, these users exhibit higher purchase intentions compared to those with lower emotional dependency.

This interaction reveals a nuanced pattern. When AI equality is low, emotionally driven users are more likely to engage with AI influencers, possibly viewing them as substitutes for social interaction or as accessible guides in complex digital environments. Conversely, when AI equality is high, the influence of emotional factors diminishes, and users across different psychological profiles behave more similarly.

The findings point to an emerging ethical concern in AI-driven marketing. Vulnerable populations, particularly those experiencing social isolation or limited digital access, may be more easily influenced by AI-generated personas designed to mimic human interaction. This raises questions about the responsible use of AI in persuasive communication and the potential for exploitation in digital ecosystems.

Privacy paradox defines consumer behavior in AI ecosystems

The finding revolves around the "privacy paradox," a phenomenon in which users express concerns about data privacy while simultaneously engaging in behaviors that expose personal information.

The research confirms that privacy concerns play a dual role in shaping consumer behavior. On one hand, users are aware of the risks associated with sharing personal data with AI systems, particularly in contexts involving virtual influencers that collect and process user preferences. On the other hand, these concerns do not necessarily reduce engagement.

Instead, the study finds that in environments where AI equality is high, users with greater privacy concerns also report higher perceived benefits of using ChatGPT. This suggests that even privacy-conscious individuals are willing to engage with AI technologies when they believe the advantages outweigh the risks.

This contradiction highlights the double-edged nature of generative AI. While it offers convenience, personalization, and efficiency, it also raises significant concerns about data security and user autonomy. The balance between these factors determines whether users ultimately adopt or resist AI-driven systems.

The study finds that privacy concerns have little impact in contexts where AI access is limited. In such scenarios, the perceived benefits of AI remain low regardless of privacy awareness, indicating that access and capability are prerequisites for meaningful engagement with AI technologies.

The findings reinforce the idea that privacy is not a standalone issue but one that interacts with broader structural and psychological factors in shaping user behavior.

ChatGPT as a general purpose technology in commerce

According to the study, ChatGPT and similar large language models are potential "general purpose technologies," a class of innovations characterized by widespread applicability, continuous improvement, and transformative economic impact.

Generative AI meets these criteria by enabling applications across multiple domains, including marketing, customer service, content creation, and data analysis. In the context of e-commerce, ChatGPT-powered systems are already being used to generate product descriptions, analyze consumer sentiment, and facilitate personalized recommendations.

The research highlights that the perceived benefits of such ubiquitous AI systems play a critical mediating role in driving consumer behavior. Users who recognize the value of AI in simplifying tasks, improving decision-making, and enhancing convenience are more likely to adopt these technologies and act on their recommendations.

However, the study also warns that the widespread adoption of AI as a general purpose technology could amplify existing inequalities if access and benefits are not distributed equitably. The concept of AI equality becomes central in this context, determining who can fully participate in and benefit from the AI-driven economy.

Implications for digital marketing and AI design

For digital marketers, building trust as a foundational element of AI-driven campaigns is important. Without trust, even the most advanced AI systems are unlikely to influence consumer behavior effectively.

Designers of AI interfaces must also account for user psychology, particularly emotional needs and perceptions of belonging. Personalization strategies that consider these factors can enhance engagement but must be implemented responsibly to avoid manipulation.

Privacy remains a critical consideration. The study suggests that transparency, data protection measures, and user control mechanisms are essential for maintaining trust in AI systems. Addressing the privacy paradox requires not only technical solutions but also clear communication about how user data is used.

The research also points out the AI divide at both local and global levels. Ensuring equitable access to AI technologies is not only a matter of fairness but also a prerequisite for maximizing the societal benefits of digital transformation.