The Perks & Perils of Empathetic AI

As artificial intelligence systems become increasingly present in customer service environments, a key question arises: How does the tone of an AI — empathetic or neutral — shape the way users experience digital interactions? And does the nature of the service itself — whether it’s emotionally driven or functional — influence how such communication is received?

This thesis explores exactly that intersection: the emotional nuances of AI communication and the influence of service context. Drawing on Social Response Theory, Expectation-Confirmation Theory, and the Hedonic-Utilitarian framework, the study aimed to investigate how AI language style affects three crucial user outcomes: perceived empathy, customer satisfaction, and reuse intention.

From Empathy to Engagement — Research Approach

To uncover these dynamics, a randomized online experiment was conducted with 183 participants, each assigned to watch one of four video scenarios. The videos simulated realistic service interactions involving an AI chatbot in either a hedonic context (travel agent) or a utilitarian context (technical support). Within each context, participants experienced either an empathetic or a neutral communication style.

Following the video, participants completed a structured questionnaire using validated scales measuring their perceived empathy, satisfaction, and behavioral intentions. The goal was to assess not only whether empathy influenced user evaluations — but also whether its effect was stronger in emotional (hedonic) services than in practical (utilitarian) ones.

What the Data Revealed — Key Findings

The results were clear and somewhat surprising. Across all conditions, empathetic communication significantly enhanced perceived empathy, satisfaction, and intention to reuse the AI system. These effects held regardless of the service context, challenging the assumption that empathy is more impactful in emotionally driven settings.

In fact, the highest scores across all three variables were found in the utilitarian-empathetic condition. This suggests that empathy may be particularly effective — and perhaps more noticeable — in function-focused services, where emotional nuance is less expected.

Additionally, the exploratory analysis revealed a rich landscape of influencing factors:

• Perceived usefulness emerged as the strongest predictor of reuse intention.
• Trust in automation had the greatest effect on satisfaction.
• Clarity and anthropomorphism (i.e., the human-like design of AI) also positively shaped the user experience.
• Younger participants and frequent AI users tended to report higher levels of satisfaction, empathy, and reuse intentions.

Interestingly, while participants explicitly prioritized functional attributes such as clarity and efficiency, empathetic cues still had a subtle but consistent emotional impact — often beneath conscious awareness. However, when empathy felt forced or overly emotional, it risked undermining trust. Some participants described such responses as artificial, even “cringeworthy,” underlining the importance of authenticity in AI design.

Practical Relevance — What This Means for AI Design

The findings offer a compelling message for designers and developers of AI services: empathetic communication works — but only when it feels natural and context-appropriate. In both hedonic and utilitarian settings, AI can benefit from subtle emotional intelligence. However, in highly functional scenarios, a credible empathic tone may have an even greater impact — precisely because users don’t expect it.

To maximise the benefits of empathy, designers should focus on combining emotional cues with high functional clarity, ensuring that AI communication meets expectations while also providing warmth and human-like interaction. Importantly, empathy should never come at the expense of clarity or credibility.

Final Thoughts

This thesis highlights not only the power of empathic AI but also its limits. When empathy feels scripted or exaggerated, it can quickly become counterproductive. But when it’s done right — balanced, subtle, and aligned with the context — it can elevate even the most mundane service interaction.

In an age where digital interfaces are becoming the new frontline of customer experience, understanding the emotional dynamics of AI is no longer a luxury — it’s a necessity.