[Paper Review] The Dilemma of Brand Campaigns in the AI Era

The introduction of AI technology in today's advertising and marketing industry has become an unstoppable trend. With the powerful weapons of cost reduction, efficiency, and hyper-personalization, AI is now reaching into the domain of creativity. However, the research Effects of AI Use Disclosure and Brand Social Value Campaign on Perceived Campaign Value, Brand Authenticity, Consumer-Brand Identification, and Purchase Intent by Kim Tae-yeon, Ko Young-ji, and Lee Jeong-hyun puts an important brake on this technological optimism. The core problem consciousness of this paper goes beyond the question "Is AI advertising effective?" to pose the more fundamental question "Where is the boundary between domains where AI may intervene and domains where it should not?"

Tool of Efficiency vs. Erosion of Authenticity

Interestingly, research results demonstrated that consumers show a dual standard toward AI — context-dependent AI backlash. In typical brand campaigns that emphasize product functionality, style, or quality, even when the fact that AI generated images is disclosed, there was no significant difference in consumer response. Consumers accept AI utilization in this context as corporate innovation or efficient tool use, and don't particularly care whether a human touch was involved. However, when brands talk about 'social values' such as diversity, equity, and inclusion (DEI), the situation changes dramatically. When AI use is revealed in socially value-oriented campaigns, value evaluation of the campaign drops sharply. This suggests that consumers perceive 'morality' and 'social correctness' as inherently human domains, not the machine's.

Trust-Building Effect of Male Virtual Humans in High-Involvement Products

The most noteworthy point among the research's key findings is the overwhelming dominance of male virtual human models in high-involvement products like laptops. Experimental results showed that in high-involvement product categories, male models elicited significantly more positive responses than female models across all indicators including advertising attitude (3.56), product attitude (3.87), and purchase intent (3.83) compared to female models (2.79, 2.75, 2.67). This suggests that the tendency for consumers to rate the expertise and reliability conveyed by masculine models more highly when encountering high-involvement products requiring high risk burden and logical information processing in purchase decision processes applies equally to virtual humans. As a result, consumers are activating the traditional frame of 'performance-centered trust' through the appearance of virtual humans.

Gender Irrelevance in Low-Involvement Products and Importance of Emotional Appeals

On the other hand, in advertisements for low-involvement products such as beverages consumed daily and repeatedly, the effect difference by model gender appeared statistically trivial or not significant. For low-involvement products, consumers react more sensitively to emotional and visual cues through peripheral routes rather than sophisticated information processing. In this context, the discriminating power of the model's gender itself weakens, and rather, sensory elements that increase visual immersion such as the model's physical attractiveness, intimacy, and styling act as the key variables determining advertising effectiveness. In some conditions, a tendency for female models to elicit somewhat more positive responses was also observed, which can be interpreted as a result consistent with the emotional approach characteristic of low-involvement products.

Strategic Persona Design: Logical Combination of Product and Model

The practical insight presented by this research lies in the point that when utilizing virtual human models, focus should be on 'Consistency (with product)' rather than reproduction of 'Realism (humanness).' For high-involvement product categories, it is effective to build masculine personas emphasizing trust and expertise to strengthen cognitive information processing, and modeling strategies should be established in the direction of strengthening brand expertise. Conversely, for low-involvement product categories, rather than fixating on a specific gender, it is necessary to plan models capable of maximizing friendliness and visual attractiveness, and to run emotional operation strategies in parallel that consider casual styling and social media utility. Ultimately, the success of virtual human marketing depends not on technical sophistication but on the sophisticated design capability to plan the optimal model persona appropriate for consumers' psychological information processing paths.