Social Media Data를 이용한 Uncanny Valley 현상의 언어권 별 차이 분석

Abstract

Objective: The present study investigated different perceptions of the uncanny valley phenomenon according to linguistic and cultural differences and difference in user sentiment concerning android and artificial intelligence (AI). Background: The uncanny valley phenomenon is recognized as an important design factor that must be considered in robots, androids, computer graphics, and animation production fields. Method: English, Japanese, and Korean tweets on Android and AI were collected. Sentiment analysis and text analysis were performed on this social media data. Results: The results showed that English users have higher rate of negative responses to androids that have appearances similar to humans compared to Japanese users; the English users use more negative expressions (adjectives). Conclusion: This implies that uncanny valley phenomenon can vary according to linguistic and cultural differences in participants and that these linguistic and cultural differences must be considered when researching and applying the uncanny valley phenomenon. Application: The results of this study could be used as empirical and quantitative data for developing design guidelines of human-friendly androids and CG contents that overcome uncanny valley phenomenon.