Abstract

The rapid rise of Generative AI (GenAI), particularly LLMs, poses concernsfor journalistic integrity and authorship. This study examines AI-generatedcontent across over 40,000 news articles from major, local, and college newsmedia, in various media formats. Using three advanced AI-text detectors (e.g.,Binoculars, Fast-Detect GPT, and GPTZero), we find substantial increase ofGenAI use in recent years, especially in local and college news. Sentence-levelanalysis reveals LLMs are often used in the introduction of news, whileconclusions usually written manually. Linguistic analysis shows GenAI boostsword richness and readability but lowers formality, leading to more uniformwriting styles, particularly in local media.

 

Quick Read (beta)