Researchers built a tool to measure how state TV channels frame the same war differently on short videos

A team created a system that reads transcripts, analyzes sentiment, and identifies visual patterns across 2,300+ YouTube Shorts about the Israel-Hamas war from major international broadcasters. The system is cheap to run and works better than expensive large AI models, which means anyone can now systematically measure how different outlets frame the same events in short-form video — the format where most people now consume news.

For years, studying how geopolitical events get framed in short-form video has been essentially impossible at scale. You could watch clips, but you couldn't systematically compare how state TV in different countries covers the same war differently, or how the framing shifted over time. This tool opens that up. It matters because YouTube Shorts and TikTok are where news actually happens now, not where reporting about news gets recycled — and we've had almost no way to study how those platforms shape what people see about major events.