Exploring Public Sentiment Toward Artificial Intelligence Apps: A Case Study of ChatGPT, Gemini, and DeepSeek in Google Apps

Abstract

Introduction: Artificial intelligence (AI) has witnessed rapid advancements in recent decades, impacting various sectors such as business, education, and entertainment. AI-based applications have become integral to daily interactions, with platforms like Google hosting popular applications such as ChatGPT, Gemini, and DeepSeek. These AI applications offer distinct approaches to technology but have the potential to influence public sentiment toward AI broadly. However, public perception remains diverse, with some embracing AI for its potential, while others express concerns regarding its implications, such as job displacement and privacy issues.

Objectives: This study aims to explore the factors that shape public sentiment toward three AI applications—ChatGPT, Gemini, and DeepSeek. Specifically, it addresses the following research questions: (1) What factors influence public sentiment toward these AI applications? (2) How do the sentiments differ between these applications? (3) To what extent is public sentiment reflective of broader perceptions of AI technology?

Methods: The research employs a case study approach, collecting user reviews from Google Play Store for ChatGPT, Gemini, and DeepSeek. Data preprocessing includes removing null entries, normalizing text, and performing tokenization. Sentiment classification is conducted using the Ekman’s Six Basic Emotions model, and sentiment analysis is enhanced using machine learning models, specifically Naive Bayes (NB) and Logistic Regression (LR). The models’ performance is evaluated based on AUC, Classification Accuracy (CA), F1 Score, Precision, Recall, and Matthews Correlation Coefficient (MCC).

Results: The analysis reveals that User Interaction and App Performance are the primary factors influencing public sentiment. ChatGPT receives the highest level of positive sentiment, particularly for its interactive capabilities. While Gemini also receives favorable reviews, its focus on intelligent search results in slightly less positive sentiment compared to ChatGPT. DeepSeek displays a more mixed sentiment, with some users appreciating its depth in data analysis, but many expressing dissatisfaction with its user interaction. Sentiment analysis further demonstrates that Joy and Surprise were the dominant emotions for ChatGPT, whereas Fear and Disgust were less prevalent across all applications.

Conclusions: This study concludes that user interaction and performance significantly drive public sentiment toward AI applications. While concerns over security and privacy exist, they are less influential compared to the experience users have with the application's functionality. The findings highlight the importance of enhancing user experience and performance for AI adoption. Additionally, the research provides insights into the need for further transparency regarding data privacy and the ethical use of AI.