When AI ethics fail and trigger negative spillovers on customer firms: evidence from event study and fsQCA

Purpose This study examines the impact of artificial intelligence (AI) ethics incidents, events in which the development or use of AI technologies violates accepted ethical norms, on publicly listed firms in the US and investigates whether these effects spill over to their customer firms within supply chains. Design/methodology/approach A dataset of 181 AI ethics incidents involving 205 US-listed firms (2018–2024) was analysed. An event study assessed stock market reactions and potential spillovers to customers, considering both intentional/unintentional incidents and six AI ethics dimensions, while fuzzy-set qualitative comparative analysis (fsQCA) explored how deficiencies across six AI ethics dimensions shape negative outcomes. Findings The event study reveals that AI ethics incidents lead to a significant decline in firms' stock returns and provoke adverse spillover effects on customers. The market reactions were found to be heterogeneous: unintentional incidents generated stronger negative reactions than intentional ones. In the case of intentional incidents, the absence of accountability emerged as a core condition in configurations associated with customer-firm losses, whereas for unintentional incidents, concurrent deficiencies in human well-being and privacy were sufficient to explain the declines, as indicated by the fsQCA results. Originality/value This study advances legitimacy theory by extending its application to AI ethics within supply chain contexts through the integration of event study and fsQCA. It provides valuable insights into the financial implications of AI ethics failures and offers practical guidance for firms, investors and policymakers seeking to anticipate, manage and mitigate the repercussions of such incidents.