Affective computing (emotion inference)
Inferring a person's emotional state and intentions from face, voice, language, physiology and context accurately enough to respond appropriately or detect deception.
State of the art
LLM-based multimodal assistants respond plausibly to user affect in conversation (2025–2026), but validated individual-level emotion inference from biometrics does not exist.
- Same expression maps to different emotions across contexts, cultures and individuals
- No ground-truth labels for internal emotional states
- Models learn demographic stereotypes from biased labels
Who’s working on it
- Rosalind Picard ↗ResearcherMIT Media Lab Affective Computing group
Founded the field (1995 book) and still leads wearable affect and stress-sensing research in 2025; co-founded Affectiva and Empatica
- Hume AI ↗Startup
Ships an empathic voice interface (EVI 3 API, July 2025) that reads vocal expression and adapts its responses
- Smart Eye (Affectiva) ↗Company
Acquired Affectiva in 2021; sells facial-expression emotion sensing for in-car driver monitoring and market research, demoed at CES 2026
- EmpaticaCompany
Picard spinout whose wrist sensor became the first FDA-cleared AI smartwatch, for seizure detection, using physiological arousal signals
- Lisa Feldman BarrettResearcherNortheastern University
Led a 2019 review showing facial movements do not reliably reveal emotions, the main scientific challenge to expression-based emotion AI
- EU AI Act emotion-recognition banProgram
Since February 2025 the Act bans emotion recognition in workplaces and schools, which limits where the technology can be deployed in the EU
What to invest in
| Kind | Investment | Scale |
|---|---|---|
| research | Longitudinal datasets pairing multimodal signals with experience-sampled self-reports | $$ $10M – $100M |
| research | Context-aware individual-calibrated affect models with validated accuracy metrics | $$ $10M – $100M |
Depends on
Required by
No other capability builds on this directly.