Comparative studies of AI-Generated and Human-Written News: The Case of NewsGPT and The Guardian

This study compares news produced by AI (NewsGPT) and humans (The Guardian), analyzing differences in news production. Using quantitative methods like topic modeling (LDA) and qualitative tools like discourse analysis and word cloud visualization, the research investigates distinctions between artificial intelligence and human journalism. The study analyzes over 100,000 news articles published between February and September 2025 across four dimensions: topic distribution, framing, language, and algorithmic bias. NewsGPT shows high topic concentration, shallow narrative depth, and omissions regarding gender, racial, and political diversity, reflecting structural biases in its data and logic. By contrast, The Guardian offers broader topics, richer emotional language, and stronger social context sensitivity. The results make clear: AI-generated news is not a mere imitation of human journalism. It carves its own path, shaped by the drive for algorithmic efficiency and the limits of its data. This distinction not only sparks theoretical questions about AI’s role as a media actor but also urges greater transparency, editorial accountability, and ethical oversight as automated reporting grows in influence.