Underreported in the US: new site uses open source LLMs to find underovered news (pressaudit.org)

A new platform called PressAudit has been launched to address the issue of underreported news in the United States by utilizing open-source large language models (LLMs). This innovative site tracks coverage trends across major news outlets, identifying gaps in reporting and highlighting stories that deserve more attention. The system employs rule-based algorithms to analyze news coverage patterns, which serve as a foundation for comparison rather than definitive judgments of coverage intent or accuracy. The significance of PressAudit lies in its potential to enhance media literacy and accountability by revealing blind spots in mainstream reporting. Users can explore a coverage heatmap that shows the disparity of reporting across outlets, view stories with divergent headlines side by side, and follow the latest coverage trends. Additionally, PressAudit emphasizes transparency by mapping out newsrooms based on their editorial leanings and public-interest quality. By fostering a more equitable news ecosystem, PressAudit not only empowers readers to make informed choices about their news sources but also encourages media outlets to improve their coverage diversity.