The Europe Media Monitor (EMM), created by the Joint Research Centre (JRC), is a robust tool that monitors over 20,000 websites and processes approximately 500,000 online pages daily in 80 languages. It is invaluable for scientific research at the JRC, allowing researchers to develop advanced language analysis methods for trend identification, text categorisation, and key entity detection. Over time, research results in these areas have been refined and integrated into the EMM system.

Advances in large language models (LLMs) and other AI tools have greatly enhanced these processes. LLMs can process and generate human-like text, making them effective for handling complex, multilingual content and tasks such as sentiment analysis and topic classification.

Our development of applications aims to improve media analysis tools, focusing on understanding online discourse, particularly in political intelligence and misinformation detection. Enhanced by LLM capabilities, our tools support diverse applications like topic classification and identifying biased text.

Current research targets the detection of persuasion techniques, misinformation, conspiracies, and polarised content, extending EMM's capabilities in categorisation and trend-spotting. We have developed an online demonstrator to showcase these methods and organise research challenges to improve our tools. A novel cross-lingual annotation measure has also been introduced to refine datasets and enhance classifier accuracy.

With the support of LLM advances, EMM and related applications provide critical insights for European institutions, assisting media monitoring and offering vital information on issues such as natural disasters, health threats, and scientific developments. These tools help researchers and policymakers stay informed about emerging issues, ultimately supporting decision-making and policy shaping, highlighting the importance of advanced language processing in today's data-rich world.

Back    Use tool