The intrinsic mechanism of public technology acceptance, risk perception, and pharmaceutical safety trust in the context of AI-driven pharmaceutical regulation

Artificial intelligence is being deeply embedded throughout the entire chain of pharmaceutical regulation. However, technological efficacy does not automatically translate into public trust. Existing technology acceptance models default to the causal sequence in which risk and trust determine acceptance, yet they overlook the possibility that the public may adjust risk perceptions in ways that are consistent with their prior attitudes toward the technology. Moreover, the over-aggregation of risk perception obscures the heterogeneous nature of different risk types. Grounded in cognitive consistency theory, this study developed a theoretical model in which technology acceptance is associated with three-dimensional risk perceptions (data privacy risk, algorithmic error risk, accountability risk), which in turn are associated with pharmaceutical safety trust. The aim was to examine whether the data are consistent with the proposed pathway running from acceptance to risk to trust. A survey was conducted among Chinese adult citizens through an online platform, yielding 412 valid responses (48.1% male, 51.9% female). Structural equation modeling was employed to test the hypothesized relationships. Technology acceptance showed significant negative associations with data privacy risk (β = − 0.39, p < 0.001), algorithmic error risk (β = − 0.35, p < 0.001), and accountability risk (β = − 0.43, p < 0.001), with the largest numerical coefficient observed for accountability risk. All three risk dimensions were significantly and negatively associated with pharmaceutical safety trust (β_privacy = − 0.28, β_algorithm = − 0.24, β_accountability = − 0.33, all p < 0.001), with accountability risk showing the numerically strongest association. The model demonstrated good fit (χ 2 /df = 2.42, CFI = 0.93, RMSEA = 0.059). The results are consistent with the theoretically proposed pathway whereby technology acceptance is associated with lower risk perceptions and higher trust, and they offer an alternative perspective on the unidirectional causal assumption embedded in classical technology acceptance models in AI governance contexts. Accountability risk emerged as a particularly strong correlate of reduced public trust, suggesting that regulatory authorities should prioritize establishing clear accountability frameworks for AI-assisted regulation as a key institutional response to public concerns.