Opalite AI: Blurring the Boundaries Between Artificial and Human Intelligence
(A2026-131904)

AUTHORS

MICHAEL HAENLEIN, ESCP Business School/University of Liverpool Management School; Valentina Pitardi, Surrey Business School; Janina Garbas, ESCP Business School; Anouk Bergner, University of St. Gallen; P. K. Kannan, University of Maryland

ABSTRACT

Recently, artificial intelligence has evolved from a tool for structured tasks to a companion that increasingly mirrors relationships once experienced only with humans. Like opalite, a man-made stone that resembles something natural, AI now blurs boundaries between the artificial and the human. This special session presents four cases of such “Opalite AI.” The first paper shows that robots giving instructions are perceived as less respectful and reduce compliance unless they display interactional justice. The second demonstrates that AI replicas of real employees create stronger relational continuity than fictional personas. The third reveals how empathic language models help victims of verbal harassment feel heard and more capable of coping. The final paper shows that matching an AI’s affective or cognitive style to users’ preferences significantly increases engagement. Together, these studies reveal how man-made agents increasingly evoke human-like responses. In the first presentation (Following Instructions from a Robot: A Justice Perspective), Valentina Pitardi, Ana Valenzuela, and Noah Castelo show that consumers comply less with instructions from robots than from humans or chatbots because robots are perceived as treating people less respectfully. Across seven studies, lower interactional justice consistently drives noncompliance. However, when robots display respectful cues or use personal information, perceived justice rises and compliance improves. In the second presentation (Traveling with Janus: Fictional vs. Employee-Based Personas in AI Companions for Customer Relationship Marketing), Janina Garbas, Vitor Lima, Michael Haenlein, and Jenny van Doorn argue that AI companions modeled after real employees strengthen customer relationships more than fictional AI personas. Employee-replica AIs create a sense of relationship continuity, boosting engagement. Through netnography, fieldwork, and experiments, the study demonstrates that replicating real humans enhances trust, connection, and relational marketing outcomes. In the third presentation (Empathy on Demand: How Empathic AI Can Sale Emotional Support for Verbal Harassment), Anouk Bergner, Philipp Winder, and Christian Hildebrand provide evidence that large language models can provide effective empathic support to victims of verbal harassment. By using linguistic markers of emotional validation, perspective-taking, and action orientation, AI responses make users feel heard and more capable of coping. The work positions empathic AI as an accessible complement to traditional support. In the final presentation (Destiny by Design: Affective versus Cognitive AI Communication Style and User Engagement in AI Fortune Telling Apps), Hyoryung Nam, PK Kannan, Jeonghye Choi, and Yoelib Kim show that aligning an AI’s communication style with a user’s affective or cognitive preference significantly boosts engagement in a fortune-telling app. A four-week field experiment reveals that higher session duration, interaction, satisfaction, and trust occur when the style matches. Stylistic fit emerges as a key driver of authentic, engaging AI experiences.