The beauty and fashion feeds of Ghanaian youth are changing. Alongside human influencers, you now find perfectly curated faces that do not exist in real life – AI influencers who recommend foundations, wigs, and outfits with algorithmic precision.
Their rise is not accidental. Global beauty brands are turning to AI because they are cheap to manage, immune to scandal, and available 24 hours a day. But technology alone does not persuade people to spend money.
Persuasion is cultural. That was the starting point of our study, recently published in Frontiers in Communication. We wanted to understand what actually makes a young Ghanaian believe, connect with, and ultimately buy on the recommendation of an artificial being.
We surveyed 817 active social media users in Ghana between 18 and 45 years, all of whom had purchased beauty or fashion products in the previous six months. The results complicate the hype around AI.
Credibility is judged through a cultural lens
Conventional wisdom assumes that people trust AI because it is technologically superior – more accurate, more data-driven. Our evidence points elsewhere.
In Ghana, a collectivist society where social harmony, respect for norms, and group endorsement shape behavior, credibility is not granted on technical competence alone. An AI influencer is seen as credible when it mirrors the cultural orientation of its audience.
When an AI uses familiar language cues, reflects local standards of beauty, and behaves in ways that align with communal values, young consumers rate it as more expert and more honest. Cultural values predicted AI credibility significantly.
This suggests that legitimacy for non-human agents is not imported; it is negotiated locally. It challenges the assumption that global AI models can be deployed in Africa without cultural translation. What reads as attractive or trustworthy in a Western market may read as alien or even deceptive here.
The critical gap between knowing and trusting
The most revealing insight from the data is a distinction most marketing campaigns ignore: credibility and trust are not the same. Credibility is a cognitive judgment – “this source knows its subject.” Trust is a willingness to be vulnerable – “I will let this source guide my decision.
“We found a very strong link between the two. AI credibility explained 59% of variance in trust. Yet credibility on its own had no direct, significant effect on purchase decision confidence. In fact, without trust, its effect turned negative. This is an indirect-only mediation.
The pathway from AI to buying confidence works only when trust is present. This matters because many brands are investing heavily in making AI look more human, more beautiful, more expert, while under-investing in the mechanisms that actually produce trust: transparency, accountability, and consistency.
For a young Ghanaian consumer who knows that an AI has no lived experience, no moral consequence, and no human to hold accountable if a product damages her skin, expertise is insufficient. She withholds confidence until trust is earned. That is a rational response, not digital illiteracy.
Emotional attachment is real, even to non-humans
Another finding that should not be romanticized: young people do form emotional bonds with AI influencers. Both cultural alignment and credibility predicted attachment.
This is not irrational. In markets where face-to-face interaction with brands is limited and physical stores are concentrated in urban centers; AI influencers fill a relational gap. They respond instantly, remember preferences, and offer constant affirmation. But we must be critical here.
Attachment to an entity that simulates care without feeling it raises ethical questions. Our study measured attachment as an outcome, not as a predictor of loyalty. Future work needs to ask whether such attachment leads to long-term empowerment or to dependency on curated realities that narrow beauty standards rather than expand them.
What this means beyond marketing
Three implications stand out. First, for brands, the playbook must change. Cultural grounding cannot be cosmetic. It is not about adding a Ghanaian flag to a post. It is about narrative, values, and visual codes that reduce uncertainty. Our model showed cultural values directly increased decision confidence, not just through credibility. When consumers see their world reflected, they decide with less anxiety. Second, trust must be engineered deliberately.
That means clear disclosure that the influencer is AI, coherent messaging across platforms, and visible human oversight. Trust is not built by pretending the AI is human, but by being honest that it is not, and showing why it is still worth relying on.
Third, for policymakers and researchers, this is a reminder that emerging markets are not late adopters waiting for Western technology to arrive. They are distinct contexts that reshape how technology is perceived. Ghanaian youth are not rejecting AI; they are filtering it through culture.
This extends Source Credibility Theory, which has long focused on human endorsers in individualistic societies, to non-human agents in collectivist settings. The study has limits. It is Ghana-specific, youth-focused, and centered on beauty and fashion – a sector where emotion drives purchase more than in, say, banking. Cross-cultural replication is essential.
Yet the broader message holds: in the rush to automate influence, we cannot automate culture. Even when the influencer is artificial, the decision to trust remains deeply human, deeply social, and in Ghana, deeply cultural.
The writer is a lecturer at University of Professional Studies, Marketing Department
Dr. Ebenezer Arthur Duncan
