Emotion-aware chatbot with cultural adaptation for mitigating work-related stress

Shi Hui Ng, Lay-Ki Soon, Tin Tin Su

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review


The accessibility to affordable and yet effective mental health support is limited due to various barriers. Given the proliferation of technology, chatbots for mental health support has been widely used. Being mindful of the users' cultural background and the ability to respond with empathy are perceived as important factors that contribute to the usability and effective communication with chatbots. Nonetheless, cultural adaptation and emotional sensitivity in mental health chatbots are not thoroughly investigated. Hence, this work aims to design and implement an emotion-aware chatbot which incorporates cultural-adaptation that could provide effective Cognitive Behavioural Therapy (CBT) interventions to Malaysian community. The emotion detection model was developed using BERT and achieved an accuracy of 0.89. For cultural adaptation, besides localised contents, Google Cloud Translation API was used as the machine translation model between Malay to English. A user study was then carried out to assess the effectiveness of emotion sensitivity and cultural adaptation in CBT-based mental health support. The ablation study shows that CBT, cultural adaptation and emotional sensitivity have positive impact on the effectiveness and usability of mental health chatbots.

Original languageEnglish
Title of host publicationAsian HCI Symposium'23 Proceedings
EditorsEunice Sari, Masitah Ghazali, Adi B. Tedjasaputra
Place of PublicationNew York NY USA
PublisherAssociation for Computing Machinery (ACM)
Number of pages10
ISBN (Electronic)9798400707612
Publication statusPublished - 2023
Event9th Asian HCI Symposium, Asian CHI 2023 - Virtual, Online, Indonesia
Duration: 28 Apr 2023 → …


Conference9th Asian HCI Symposium, Asian CHI 2023
CityVirtual, Online
Period28/04/23 → …


  • chatbot
  • cultural adaptation
  • emotion detection
  • Mental health intervention

Cite this