Research from the University of Birmingham, Aarhus University, and Linnaeus University has identified a bidirectional influence in human-AI customer service interactions: as conversational AI becomes more anthropomorphised through adaptive learning and empathetic engagement, customers unconsciously mirror these robot-like communicative patterns back. The mechanism operates through a feedback loop where machine learning amplifies this effect—the robot learns from customer input and adjusts its behaviour accordingly, prompting customers to internalise increasingly formalised, algorithmically-shaped communication styles. This phenomenon, termed "robotoid humanness," raises immediate questions for teams deploying sophisticated AI agents: if your Zendesk or Freshdesk automation is designed to be maximally personable and adaptive, are you inadvertently training customers to communicate in ways that reinforce algorithmic patterns rather than authentic human expression?
The ethical implications cut deeper than communication style. Repeated exposure to AI-mediated interactions can reshape customer self-perception, particularly through confirmation bias and algorithmic feedback loops. Customers may begin depending on these systems for validation, where positive reinforcement from an AI agent enhances confidence whilst negative cues erode self-esteem—a dynamic that becomes problematic at scale when millions of interactions are governed by the same underlying models. For support leaders, this surfaces a critical tension: the same personalisation and learning capabilities that drive engagement metrics and CSAT scores may be subtly constraining how customers perceive themselves and interact with your brand. The research suggests that technological readiness, cultural background, and personality traits all moderate this effect, meaning the impact is neither uniform nor predictable across your customer base.
The practical challenge for CX teams is calibrating anthropomorphism deliberately rather than by default. Rather than maximising human-like behaviour in AI agents simply because it improves short-term engagement, teams should consider whether their automation strategy accounts for the long-term cognitive and identity effects on customers. This doesn't necessarily mean deprioritising AI—it means being intentional about where and how you deploy it, and whether human handoff points, transparent communication about interacting with AI, or deliberate breaks in algorithmic feedback loops should be built into your customer journeys. The question isn't whether to use AI in customer service; it's whether your implementation acknowledges that customers are being shaped by the systems you've built to shape their experience.
Interacting with customer service AI can make people act more like robots, research suggests Digital Information World
Interacting With AI Might Make You Act 'More Robot-Like', Researchers Warn ScienceAlert
Researchers Warn Interacting With Customer Service AI Could Cause 'Robotoid Humanness' International Business Times UK