Artificial intelligence in service: a commentary on AI-expressed emotions and customer’s trust dilemma
Sarbani Mishra, Poonam Kumar, Saumyaranjan Sahoo, Raj Agnihotri · The Journal of Marketing Theory and Practice · 2026
Conversational AI can now perform emotionally charged frontline work, yet it is not clear if algorithmically produced emotional signals build or erode customer trust. We explore AI-expressed emotions through the lens of the customer’s trust dilemma in service delivery. Emotional signals can automatically raise affective trust, whereas cognitive signals raise cognitive trust only when the expression is judged appropriate. The same empathic signal can deepen one form of trust while leaving the other or even undermining it. Drawing on the Emotions as Social Information model, a dual-process account, and the cognitive/affective trust distinction, we deliver a snapshot of current understanding across five domains: the constructs and mechanisms of AI-expressed emotions, the trust dilemma itself, its boundary conditions, its ethical and transparency implications, and the emerging contexts that will reshape it. Simultaneously, we delineate open questions for future research within each domain. This commentary offers a nuanced take on AI-expressed emotions in services, examining its promise and perils for firms, customers, and regulators, and serves as a catalyst for scholarly and practical interest.