Global news media coverage of artificial intelligence (AI): A comparative analysis of frames, sentiments, and trends across 12 countries

Abstract

This study examined the coverage of artificial intelligence (AI) in newspapers from 12 countries by analyzing news articles (N = 38,787) collected from 12 mainstream English newspapers, between 2010 and 2023. We used LDA topic modeling to identify prevalent frames in the news articles and SentiStrength to examine sentiments in the news headlines. Framing theory was applied in interpreting our results. Our analysis identified nine frames across newspapers: AI impacts on businesses, economy, and jobs (37.40 %), AI transformations in education and research (17.70 %), AI in national security and global partnerships (11.20 %), AI disruptions in media and creative industries (9.6 %), AI-based innovative solutions (7.30 %), AI regulations, ethics, and data privacy (6.40 %), AI competition and market dynamics in tech industries (4.90 %), AI in healthcare and climate change (3.47 %), and AI in politics, elections, and public opinion (2.03 %). A comparative analysis suggested that the Global North newspapers gave relatively lower coverage to AI-based innovative solutions and AI in healthcare and climate change while AI regulations, ethics, and data privacy and AI disruptions in media and creative industries received minimal coverage from the Global South newspapers. Our overall sentiment analysis indicated that 21.04 % of news headlines evoked negative, 13.33 % positive, and 65.63 % neutral sentiments. The Global North newspapers such as The Guardian and The NYT framed AI negatively in the 24 % of their news headlines, while the Global South newspapers such as China Daily and Bangkok Post framed AI positively in the 14.5 % of their news headlines.