Emotion AI: Utopian Promises, Dystopian Realities

Emotion AI: Utopian Promises, Dystopian Realities

When: December 10th, 16:30-18:00, followed by Apéro

Where: Digital Society Initiative, Rämistrasse 69, Zurich, or online via Zoom link

Organizers: DSI ethics and AI&Law communities

Registration: Please register here

Description

Emotion AI, a branch of affective computing, promises to algorithmically detect, infer, or predict people’s emotions, moods, and mental states. These systems are rapidly spreading, from entertainment and consumer technologies to education, healthcare, and the workplace—despite an absence of regulation in the United States and weak regulatory efforts in the European Union. Drawing on a series of empirical studies with diverse social groups as well as critical analyses of emotion-AI technologies, scholarly narratives, and patents, this talk examines how emotion AI amplifies existing inequities and produces new forms of harm. Rather than promoting well-being, fairness, or improved conditions, emotion AI disproportionately burdens marginalized communities, deepens identity-based vulnerabilities, and exacerbates the very structural challenges its proponents claim it will remedy. I argue that emotion AI is not merely a technical innovation; it is a sociotechnical project rooted in power, profit, governance, and the regulation of emotional life. Its deployment relies on pervasive surveillance, coerced emotional labor, and the systematic absence of meaningful consent within unequal power relations. I conclude by discussing why emotion data must be recognized and treated as highly sensitive in research, practice, and policy; why technical fixes cannot resolve the core harms of emotion AI; and what more just alternatives might look like for workers and other data subjects navigating increasingly, and often dubiously, “emotion-aware” sociotechnical environments.