A recent study explores how far advertisers should go in personalizing advertisements using AI-generated faces. While using AI to make ads appear more personal can boost their effectiveness, the research shows there is a limit. Once the AI-generated face becomes too similar to the viewer, it can actually reduce the desire to buy. This discovery highlights the fine line between creating a sense of connection and crossing into discomfort. Generative artificial intelligence (AI) has given advertisers powerful new tools. Rather than just creating generic ads for broad audiences, companies can now tailor content to individuals. For example, AI can generate images of a consumer wearing different outfits or even create a face that looks like them. This shift marks the rise of "synthetic advertising," where AI-generated or manipulated content is used to influence consumers. While this allows for more personalized experiences, it also raises questions about how much personalization is too much. To understand the impact of facial similarity, researchers used a method based on facial recognition technology. They converted faces into digital representations and compared them using an algorithm. This allowed them to measure how similar two faces were on a scale from 0 to 1. When tested against human judgment, the algorithm proved highly accurate in determining facial similarity. This made it possible to conduct experiments without needing new participants for each comparison. In two separate experiments, participants viewed AI-generated advertisements featuring faces that varied in similarity to their own. The results showed a clear pattern: as facial similarity increased from low to moderate levels, purchase intention also increased. However, once the similarity became very high, the effect reversed. Participants became less likely to buy the product. This "inverted U" pattern suggests that there is an optimal point for personalization—around 57.6% facial similarity—beyond which the ad becomes less effective. This phenomenon may be due to several factors. One is the feeling of intrusion: if an ad seems too personal, it can make consumers uncomfortable, thinking the brand knows too much about them. Another is the "uncanny valley" effect, where almost-human but slightly unnatural faces can cause unease. Additionally, when personalization is too obvious, it can make consumers more suspicious of the brand’s intent. While AI-generated faces can create a sense of closeness, they must be used carefully to avoid crossing into awkwardness or manipulation. The study does not argue against using AI in advertising, but rather emphasizes the need for balance. Advertisers should aim for a level of personalization that feels natural and engaging, without making consumers feel invaded or manipulated. As the technology continues to evolve, the challenge will be to find the right balance between personalization and respect for consumer boundaries. The key question isn’t just how much an ad can be tailored, but when that tailoring becomes intrusive instead of helpful.