Artificial intelligence systems now deliver market‑facing promises via chatbots, recommendation engines, automated decisions, and generative interfaces. As failures, misuse, and misrepresentation become more common, conventional brand‑crisis models fall short of explaining how responsibility is allocated when technical causation, customer‑facing control, and governance duties are spread across an AI system, its developer, deployer, vendor, and user.
The paper first distinguishes three states: an AI/algorithmic incident is a technical malfunction; an AI‑related organisational crisis is the operational impact on a firm; an AI‑related organisational scandal emerges when the public moralises the issue, challenging the firm’s reputation. This taxonomy provides a structured entry point for responsibility attribution.
Drawing on a structured, federated scoping synthesis of verified academic and primary sources, the proposed sociotechnical process theory comprises four linked steps: incident configuration shapes actor‑specific attribution; attribution informs assessments of capability, integrity, fairness, and relational quality; these assessments may be amplified by public moralisation, potentially escalating the incident into a scandal. The theory accounts for why algorithm involvement can sometimes buffer negative brand reactions, whereas robot or chatbot failures tend to shift blame directly onto the associated firm.
To address this complexity, the paper introduces “accountable transparency” as a response configuration. Its essential components are timely notice, an intelligible account, clear role‑responsibility acknowledgement, remedial action, evidence of correction, and accessible recourse. Together, these elements enable firms to be open while demonstrably assuming responsibility.
Empirical evidence shows that practices aligned with accountable transparency yield stronger conditional and proximal inferences about blame, trust, satisfaction, firm evaluation, and communication credibility. In contrast, claims about durable reputation, brand equity, or market performance receive weaker support.
Review