Cross-Cultural Emotion Recognition in AI: Enhancing Multimodal NLP for Empathetic Interaction

Abstract

It investigates using cross-cultural understanding of emotions and empathy to make HCI better. Using techniques such as NLP, examining text, sound, and visuals, along with transformer models, the research enables AI to identify emotions. The system was most accurate in identifying both positive and neutral emotions but struggled slightly in detecting anger or sadness. Contextual and organized answers were generated by the empathetic response module, which achieved an average of 4.3/5 in empathy metrics. There are still difficulties in evoking strong emotions in audiences, especially when it comes to portraying complex emotions. The research emphasizes that AI systems may fail to recognize certain emotions if they are not designed to detect diverse cultural expressions of emotions. Topics related to the privacy of emotional data and problems with algorithm bias are openly discussed, highlighting the need for open and responsible work on AI. Study results contribute to building AI that understands emotions, which helps users in industries such as healthcare, education, and service, and also supports cultural understanding and ethical design in AI