Generative AI adoption and barriers in the insurance industry

Abstract This study examines the adoption of Generative Artificial Intelligence (Gen AI) in the insurance industry through an institutional lens, drawing on semi‐structured interviews with 21 senior insurance professionals. While Gen AI offers significant potential for a data and text‐intensive industry, adoption remains cautious and uneven. Using qualitative thematic and content analysis, the study maps current and prospective Gen AI applications and identifies the barriers shaping their implementation. The findings show that Gen AI adoption in insurance is structured less by technical feasibility and more by institutional pressures related to regulation, professional norms, and trust. Insurers prioritize low‐risk, internally focused applications that align with dominant institutional logics of prudence, accountability, and risk control, with a continued reliance on human oversight, while higher‐stakes and customer‐facing use cases are adopted incrementally or deferred. Regulatory ambiguity emerges as a prominent constraint, encouraging strategies of anticipatory compliance and risk minimization. By situating Gen AI adoption within the institutional context of insurance, this study advances the literature on AI governance in regulated industries by showing how regulatory expectations, professional norms, and embedded risk logics actively shape where and how value from Gen AI is realized in practice. The study provides novel empirical insight into the uneven and selective nature of adoption across use cases and highlights how regulatory ambiguity may dampen beneficial innovation. In doing so, it underscores the importance of regulatory clarity, proportional risk‐based governance, and supervisory guidance in enabling responsible adoption, with implications for both policy and industry practice.