Unveiling Public Sentiment Towards ChatGPT: Sentiment and Thematic Analysis of X (formerly Twitter) Discourse
Vince Ryan Arboleda, Brian Steven Pajarillo, Louther Jan Adarle, Dawn Angelika Arsenal, Eddie De Paula · Philippine Journal of Science Engineering and Technology · 2024
As generative AI technologies like ChatGPT become increasingly integrated into various aspects of daily life, understanding public perception is crucial for guiding responsible development and ethical deployment. This study conducts a comprehensive sentiment analysis of Twitter discourse, utilizing an innovative approach that integrates Plutchik’s Wheel of Emotions, the NRC Word-Emotion Association Lexicon, and the VADER algorithm. By analyzing a dataset of 39,051 tweets, the research aims to identify predominant emotions, sentiment intensity and distribution (positive, negative, and neutral), and underlying themes within the discourse. The findings reveal that trust, anticipation, and joy are the most frequently expressed emotions, reflecting a generally positive reception of ChatGPT. Specifically, 54.4% of the tweets conveyed positive sentiments, 17.02% were negative, and 28.58% were neutral. Thematic analysis, facilitated by Latent Dirichlet Allocation (LDA) and Gibbs Sampling, uncovers key themes related to ChatGPT’s potential, functionality, and utility. This study contributes to a deeper understanding of public attitudes towards generative AI technologies, providing valuable insights for developers, policymakers, and researchers in addressing the ethical, practical, and societal implications of AI integration into everyday life.