Abstract
Artificial Intelligence (AI) is a rapidly evolving technology with far-reaching implications, and it has sparked intense public discussion on social media platforms. This study conducts a comprehensive sentiment analysis of AI-related posts on Twitter, YouTube, and Reddit from 2017 to 2024, aiming to evaluate public perceptions and attitudes toward AI. Approximately 133,004 social media posts were analysed using a fine-tuned RoBERTa model for sentiment classification, along with Latent Dirichlet Allocation (LDA), n-gram analysis, and word co-occurrence mapping for topic modelling. The findings reveal that a considerable portion of social media users express negative sentiments towards AI (46.09%), followed by neutral sentiments (39.29%) and positive sentiments (14.62%). LDA analysis identified 20 distinct topics, including AI ethics, job impacts, and philosophical implications. Temporal analysis revealed a notable increase in AI-related discourse starting in 2022, accompanied by evolving sentiment patterns. The results reveal a complex landscape of public perception regarding AI, characterised by ongoing concerns about its societal impacts, coupled with a growing interest in the AI's technical aspects. Future research could explore cross-cultural perceptions of AI and examine how direct interactions with AI systems influence public attitudes.
Original language | English |
---|---|
Publication status | Unpublished - 22 Nov 2024 |
Event | The Second UK AI Conference 2024 - University of Birmingham, Birmingham, United Kingdom Duration: 22 Nov 2024 → 22 Nov 2024 https://uk-ai.org/ukai2024/ |
Conference
Conference | The Second UK AI Conference 2024 |
---|---|
Abbreviated title | UK AI |
Country/Territory | United Kingdom |
City | Birmingham |
Period | 22/11/24 → 22/11/24 |
Internet address |