Sentiment Analysis of Public Perceptions on ChatGPT and Generative Artificial Intelligence (GenAI): Model’s Performance Evaluation and Examining Benefits and Risks in Education and Healthcare

Abstract

As GenAI systems become more integrated into daily activities, understanding how people react to these tools is critical for responsible design and governance. This study provides a large-scale, longitudinal analysis of public sentiment toward ChatGPT and GenAI by integrating transformer-based sentiment classification, temporal trend analysis, and sector-specific topic modeling for education and healthcare. Using over one million English-language posts collected between November 2022 and December 2023, we quantify sentiment patterns over time and identify domain-specific themes of perceived benefits and risks. A comparative evaluation of traditional machine-learning (ML) models (logistic regression, support vector machines, random forest), deep learning (DL) architectures (convolutional neural networks, long short-term memory), and Bidirectional Encoder Representations from Transformers (BERT) was conducted. DL models outperform classical ML, and BERT emerges as the most effective classifier, achieving 98% accuracy with a near-balanced profile across accuracy, precision, recall, and F1-score, outperforming traditional approaches. Using the best-performing model, the findings show that ChatGPT sentiment is predominantly positive, alongside a substantial minority of negative sentiments. Topic modeling reveals domain-specific benefits and risks in education and healthcare discourse. In education, ChatGPT promotes personalized learning, accessibility, and teacher support, but it also raises plagiarism, academic dishonesty, and data privacy concerns. In healthcare, GenAI improves patient information, diagnostics, and administrative efficiency, but it also raises concerns about misinformation, ethics, and empathy. Overall, the research provides evidence-based guidance for technology developers, educators, healthcare professionals, and policymakers taking advantage of GenAI while addressing its associated social and ethical issues.