Real-time monitoring of emotional workers' mental health AI...KAIST, Chung-Ang University, Akron University Research Team Develops Technology
Feb 11, 2025
It has become possible to monitor the burden felt by service workers called
'emotional workers' through artificial intelligence in real time.
Service workers, such as call center counselors, are often faced with situations in which they have to express emotions different from those they actually feel in the process of responding to customers. The psychological load that occurs in the process of controlling emotions is called 'emotional workload', and excessive workload can lead to burnout and depression.
Previously, studies were attempted to monitor emotional states to increase safety by preventing workers' work overload, but the focus was mainly on the intellectual worker's 'cognitive workload'. In addition, existing emotion detection AI (artificial intelligence) models diagnose emotions based on users' facial expressions and voices, making it difficult to measure the emotional workload of emotional workers who have to suppress their emotions and respond kindly.
Lee Eui-jin, a professor of computer science and technology at the Korea Advanced Institute of Science and Technology (KAIST), said he developed an artificial intelligence model that can monitor workers' emotional workload in real time in collaboration with Professor Park Eun-ji of Chung-Ang University and Professor James Defendoff of Akron University in the U.S.
The research team first collected data from multiple modal sensors such as voice, behavior, and bio-signals of 31 call center counselors. From the collected bio-signals, 228 features were found by extracting skin conductivity (EDA), brain waves (EEG), electrocardiogram (ECG), and body temperature that measure the electrical activity of the brain.
As a result of experimenting by learning this on an artificial intelligence model, it distinguished situations with high emotional workload and situations without emotional workload by 87% accuracy.
As a result of the study, the model's performance was rather degraded when it included a voice (such as a voice that is generally known to be useful for workload measurement). For call center counselors who need to suppress their emotions, it means that voice data may not be suitable as a classification criterion for emotional workloads. Rather, it was confirmed that characteristics such as skin conductivity and body temperature of counselors were more effective in measuring workload.
The research team plans to demonstrate the technology developed this time by applying it to a mobile app that can manage the mental health of emotional workers.
Service workers, such as call center counselors, are often faced with situations in which they have to express emotions different from those they actually feel in the process of responding to customers. The psychological load that occurs in the process of controlling emotions is called 'emotional workload', and excessive workload can lead to burnout and depression.
Previously, studies were attempted to monitor emotional states to increase safety by preventing workers' work overload, but the focus was mainly on the intellectual worker's 'cognitive workload'. In addition, existing emotion detection AI (artificial intelligence) models diagnose emotions based on users' facial expressions and voices, making it difficult to measure the emotional workload of emotional workers who have to suppress their emotions and respond kindly.
|
The research team first collected data from multiple modal sensors such as voice, behavior, and bio-signals of 31 call center counselors. From the collected bio-signals, 228 features were found by extracting skin conductivity (EDA), brain waves (EEG), electrocardiogram (ECG), and body temperature that measure the electrical activity of the brain.
As a result of experimenting by learning this on an artificial intelligence model, it distinguished situations with high emotional workload and situations without emotional workload by 87% accuracy.
|
The research team plans to demonstrate the technology developed this time by applying it to a mobile app that can manage the mental health of emotional workers.
This article was translated by Naver AI translator.
