A Longitudinal Analysis of Artificial Intelligence Coverage in Technology-Focused News Media Using Latent Dirichlet Allocation and Sentiment Analysis

Understanding media discussions on artificial intelligence (AI) is crucial for shaping policy and addressing public concerns. The purpose of this study was to understand sentiment regarding AI in the media and to discover how the discussion of topics changed over time in technology-related media outlets. The study involved three overall steps: data curation and cleaning to obtain a high-quality, timely dataset from a list of relevant technology-news-oriented websites; sentiment analysis to understand the emotion of the articles; and Latent Dirichlet Allocation (LDA) to uncover the topics of discussion. The study curated and analyzed 22,230 articles from technology-focused media outlets between the period 2006 and July 2024, split into three time periods. We found that discussion on AI-related topics has increased significantly over time, with sentiment generally positive. However, since 2022, both negative and positive sentiment proportions within articles have risen, suggesting growing emotional polarization. The introduction of ChatGPT 3.5 in November 2022 notably influenced media narratives. Machine learning remained a dominant topic, while discussion on business and investment, as well as governance and regulation, has gained prominence in recent years. This study demonstrates the impact of technological advancements on media discourse and highlights increasing emotional polarization regarding AI coverage in recent years.