Navigating The AI Disclosure Dilemma in Marketing Communication : How Managers Perceive and Make Disclosure Decisions

Sammanfattning

Artificial intelligence is becoming increasingly integrated into marketing communications, allowing businesses to create, modify, and distribute customer-facing content more efficiently. However, companies face a disclosure dilemma as consumer concerns and regulatory expectations around AI-generated content grow. While disclosing AI usage might support legitimacy and compliance, it can also damage perceptions of authenticity, trust, and brand credibility. This study explores the research question: How do managers perceive and navigate the emerging AI disclosure dilemma within organisations? Using an abductive qualitative approach based on fourteen semi-structured interviews with managers engaged in marketing, communication, branding, or similar organisational activities, the study adopts an interpretivist and constructionist perspective. By using Braun and Clarke's (2006) thematic analysis to assess the data collected from interviews, the study was able to be open to inductive findings while also being guided by Oliver’s (1991) strategic response framework and Institutional Theory. The findings indicate that managers perceive AI disclosure as an irreducible tension shaped by uneven institutional pressures, contextually influenced strategic risk, and uncertain regulatory boundaries. Six themes describe how managers construct and navigate the dilemma: the dilemma as an unavoidable reputational tension, institutional pressures as uneven and platform-driven, disclosure risk as context-dependent, responses across an avoidance–compliance–proactive adoption spectrum, the normalisation trajectory, and the proportionality problem. The study further makes a theoretical contribution by identifying platform-mediated isomorphism, extending Oliver’s framework through proactive adoption, and shifting the AI disclosure dilemma from consumer-facing experimental research to the managerial and organisational level. The findings suggest that organisations should develop clearer internal disclosure frameworks in the current phase, before Article 50 of the EU AI Act becomes fully enforceable.