We read other people. We gauge their mood, infer their feelings and try to predict responses to what we say and do. This helps humans succeed in complex societies where working together is crucial to thriving. Without conscious awareness, we continually observe each other – noticing minute changes in a person’s voice, posture, movement, facial expressions and eye contact. While we are not always accurate in understanding each other, recognising another person’s emotional state remains a critical ability in our hypersocial world.
With a new emphasis on growing self-awareness in the workplace, companies are applying established technology in new ways to detect changes in emotional states and innovating with new technologies to enable applications and devices to recognise human emotion.
Some methods of emotion detection are already well known: sentiment analysis of text articles, emails, chat and social posts are commonplace with increasing levels of accuracy. Face tracking is another well-known method of emotion detection where applications compare video data to a large database of facial expressions. As datasets increase exponentially, the algorithms processing the text or video data become more powerful.
The combination of speech technology and AI techniques allow applications to identify patterns in voice and sound data from microphones or recorded calls. Applications 'listen' for heightened or lowered arousal within a voice. This is particularly pertinent in high-risk environments such as transport where train drivers, for example, may be monitored to ensure they are alert and focused on the job.
Some applications monitor the way we interact with our devices over time, for example, our patterns of typing to infer changes in our emotional state. Devices that are worn or carried can differentiate between relaxed or agitated movements and gestures. Our patterns of breath can be detected with camera data or sensors on fabric that detect chest movement. Devices can monitor your heart, blood pressure, skin conductivity, muscular contractions, neurotransmitters and hormones. All of this data provides information about changes in your emotional state. By combining several methods of tracking emotional states, accuracy increases.