Researchers have developed a pipeline for analyzing media bias and framing in online news using large language models (LLMs). The pipeline groups articles into topics and events, adds named-entity and sentiment annotations, and compares news sources through various metrics. It was applied to over 8,000 Albanian news articles and showed moderate agreement with automated annotations. The study also compared two annotation prompts and found that a simpler prompt produced more consistent results but reduced execution time. --- Why it matters: This work matters to AI researchers because it demonstrates the potential of LLMs for media bias analysis, which can be useful in settings where manually verified datasets or specialized language tools are limited. Source: https://arxiv.org/abs/2608.17454