Affective computing

Affective computing is the interdisciplinary field that studies and develops systems and devices that can recognize, interpret, process, and simulate human affects—emotions, moods, and physiological states. Coined by Rosalind Picard at MIT in 1997, the field draws on computer science, psychology, cognitive science, and neuroscience to build machines that can respond to human emotional expression. Applications range from empathetic virtual assistants and mental health monitoring to lie detection and sentiment analysis of social media.

The central methodological challenge is the grounding problem: machine learning models trained on facial expressions, vocal patterns, or physiological signals learn correlations, not causal relationships, and these correlations are culturally specific, context-dependent, and historically variable. A smile means different things in different cultures; a raised heart rate can signal fear, excitement, or physical exertion. The risk is that affective computing systems impose a universal emotional taxonomy on culturally diverse populations, producing what critics call emotional