Social Perception of Artificial Intelligence on Twitter: A Comparative Study on Global South and Global North Countries

Abstract

This study examines the public perception of ChatGPT on Twitter, comparing approaches from Global South (India and Turkey) and Global North (United Kingdom and United States) countries. It utilizes a text-as-data approach, applying sentiment analysis, text analysis, and topic modeling techniques on tweets in English containing “ChatGPT” collected between January 1 and June 1, 2023. The VADER sentiment dictionary is used for sentiment analysis, while n-grams and BERTopic are employed for text analysis and topic modeling, respectively. Sentiment analysis reveals a predominantly positive perception of ChatGPT across both regions, with some concerns expressed regarding potential negative impacts. Text analysis and topic modeling indicate that discussions revolve around major technology companies and leading figures, with the Global North focusing more on sectoral applications and technical aspects, while the Global South evaluates ChatGPT within local language and cultural contexts. The findings demonstrate that socio-cultural differences and technological development levels between regions are reflected in the social perception of ChatGPT. The study contributes to understanding the early reception and perception of generative AI and highlights the importance of addressing privacy concerns, developing culturally compatible solutions, and examining the potential impacts of ChatGPT in various domains.