Hybridization of metaheuristics and NLP approach to examine public opinion towards virtual voice assistants

Virtual voice assistants have become ubiquitous in our modern lives and continue to evolve. With the proliferation of virtual assistants such as Alexa, Cortana, and Siri in households, it is important to go beyond the potential benefits and explore the factors that may lead to user anxiety and trauma. This study examines posts on social media to analyze sentiment toward virtual voice assistants and identify areas of concern that directly impact their use and potential to cause trauma or anxiety. To conduct this study, we collected and processed 13,578 Instagram, Facebook and X (Previously Twitter) posts using sentiment analysis and topic modeling with Python libraries. In addition, K-means clustering and particle swarm optimization (PSO) analysis were performed. The results show that the majority of posts (60.32%) express positive sentiments toward virtual voice assistants, while a significant percentage reflect negative or neutral sentiments. We identified ten major issues of concern related to the use of virtual assistants, including privacy and security, technology addiction and anxiety, cyberbullying, and technology-based trauma therapy. The scope for future research on virtual assistants is broad and includes areas such as privacy, ethical implications, and strategies to improve usability.