New tool aims to revolutionise the media monitoring market
In today’s digital world, the huge number of news sources is a challenge for media monitoring companies. But according to the EU-funded project MONITIO, this can be tackled by using artificial intelligence to discover, analyse and simplify media monitoring beyond the human scale. “MONITIO’s approach to a revolutionary media monitoring platform includes following responsible AI principles and keeping humans in the loop, while providing powerful analysis, inference and visualisation tools only possible through AI,” argues Carlos Amaral, MONITIO project coordinator and CEO at Priberam. The concept of the platform was first designed in the EU-funded project SUMMA, in which Priberam was one of the partners. The company then used the research results to reimagine the media monitoring process and build a new product powered by AI.
The roles of AI
Artificial intelligence was the basis of two distinct and complementary functions of MONITIO. The first is automation – necessary to address the matter of media monitoring at a global scale, considering the volume of content and number of languages. In fact, when the platform started to receive more than 150 000 articles, the team reorchestrated the ingesting and enrichment pipeline and adopted a new search engine. The second role is to assist users in their monitoring tasks and help them extract meaningful insights from the immense volume of content. “The whole content enrichment pipeline is based on AI technologies: topics classification, named entities recognition and disambiguation, single and multi-document summarisation, translations, articles clustering, and video and audio content transcription,” lists Amaral.
MONITIO’s accomplishments
At first, the media monitoring tool operated in five languages, and now all features are available in more than 30 languages. Project partner University of Cambridge was responsible for researching the fact-checking functionality of the platform. The content verification feature allows media professionals to validate news and statements through AI modules that generate customised questions to verify claims, and through the tool’s broad coverage of sources. “This module was evaluated by journalists who are fact-checking specialists. They were able to track the evolution of news stories and check facts assisted by MONITIO’s question generation model,” says Amaral. The evaluation showed the potential uses of the feature and provided recommendations for future improvements. Another important achievement for MONITIO was the curation models that can learn and improve automatically with user feedback. Amaral noted that a large language model (LLM) was recently integrated in the platform to explore conversational ways of interacting with the content. It is, for instance, capable of explaining why a certain topic is trending or what the connection between two news subjects is.
Ready for the market
During the research, MONITIO had to overcome challenges related to access to updated media content, GDPR compliance, scalability and user experience. Now, the platform is available for the Portuguese and Spanish markets, where Priberam already has content license agreements with national collective copyright organisations. It has also secured content in the UK, although the region is not a current focus for MONITIO. Priberam is also considering licensing the platform in a white label model for media monitoring companies that already have content in different countries.
Keywords
MONITIO, AI, media monitoring, artificial intelligence, media, content, journalists, fact check, fake news
