Affectiva, now part of Smart Eye, works on Emotion AI – inferring emotional and cognitive states from facial expression and behaviour. It originated in MIT Media Lab research, and unusually for the field it publishes a science overview rather than asking you to take the capability on trust.
Three applications are named. Media Analytics covers ad testing, entertainment content testing and qualitative research, where participants consent to being recorded while they watch. In-Cabin Sensing AI applies the same technology to drivers – detecting drowsiness and distraction, which is a safety application with a clear rationale. An Emotion SDK supports academic behavioural research.
Two things to weigh honestly. Inferring emotion from facial expression is scientifically contested – expressions map to internal states inconsistently across people and cultures, and the research community is not settled on it. And any deployment involves recording faces, so informed consent and data handling are not optional extras. Affectiva’s own consented research and automotive-safety framing is the right way to use this; treat claims of reading emotion at scale from unconsenting subjects with far more scepticism.
