Ever notice that when a product looks "too perfect" you might start doubting it? A recent study explored how AI labels shape young shoppers' perceptions. Participants viewed three conditions on an e-commerce site: an AI disclosure, a human creator label, or no label. For utilitarian items (like tools or appliances), the AI label lowered feelings of authenticity, visual appeal, and social presence compared to no label. These downsides disappeared for hedonic products (like fashion or accessories). Visual appeal and social connection then boosted both purchase intent and brand trust. Surprisingly, the AI label also directly increased purchase intent, though the total effect was weak due to these opposing paths. The findings call for context‑aware policies and clear guidelines on AI‑generated product images in online stores.
The experiment employed a 3 × 2 between‑subjects design (disclosure condition × product type) on a Taobao‑like platform. Participants were randomly assigned to AI label, human label, or no‑label conditions. Products were categorized as utilitarian (functional, low‑involvement) versus hedonic (fashionable, high‑involvement). A moderated‑mediation framework with bootstrapped confidence intervals was used to test causal pathways. Findings show that AI disclosure lowered perceived authenticity, aesthetic appeal, and social presence, but only for utilitarian items. These mediated paths positively predicted purchase intention and brand trust. Independently, the AI label had a direct positive effect on purchase intention (b = 0.308, p = 0.010) regardless of product type, while showing no direct impact on brand trust. Because the direct positive effect opposes the indirect negative pathways, the overall effect appears weak or even negative—a case of inconsistent mediation (suppression). The study suggests that source‑information consistency alone does not explain algorithm aversion and that platform‑level governance must be context‑sensitive, considering product type and visual presentation.
