When Emotion AI Meets Strategic Users.

The article focuses on the adoption and economic value of emotion AI—artificial intelligence systems that detect and respond to human emotional expressions—in customer care settings involving strategic users who may misrepresent their emotions to gain more resources. Using a game-theoretical model, the study finds that emotion AI is valuable to firms when the negative spillover effects of emotional expressions (such as public displays of dissatisfaction harming reputation) are relatively small compared to losses from misallocating resources. The research highlights that algorithmic noise in emotion detection does not necessarily undermine AI's value and can sometimes improve social welfare by reducing strategic emotional escalation. It further compares AI and human service systems, showing that AI can outperform humans in recognizing emotions and allocating resources consistently, and that AI monitoring of employees may be more profitable than human monitoring, especially when some noise is present. The findings emphasize the need for carefully designed allocation policies and regulatory oversight to balance firm profits, user welfare, and societal outcomes in emotion-driven AI applications.