Very soon, CTIC will begin leading new work within the BRAVE-WOW project focused on analysing how gender-based and harassment are discussed online, particularly in relation to the world of work. The activity will run for 20 months, bringing together expertise in language, data ethics and social impact.
This work will explore how people express their experiences, emotions or concerns in different languages and cultural contexts. The goal is to better understand the invisible barriers that prevent people from reporting workplace violence such as fear, shame, or lack of trust and to support early intervention efforts.
To do this, the NLP team will analyse publicly available and anonymised content from social platforms. The work is divided into five main phases:
- Defining the framework: Reviewing good practices and selecting languages, tools and data sources, with full respect for privacy and ethics.
- Assessing platforms: Evaluating where and how social conversations about GBV and work occur, and which platforms are most suitable for study.
- Collecting and processing data: Using multilingual methods to safely process and clean online content.
- Analysing sentiment and causality: Looking at emotional tone and possible links between what people say online and real-world experiences.
- Sharing insights: Summarising trends to support the development of early detection tools in future phases of BRAVE-WOW.
This approach builds on recent examples in Europe. In 2023, Italy’s national statistics institute (ISTAT), together with UNECE, explored how online conversations about femicide and violence against women could be compared to official helpline data, providing new insights into when and how people speak out online.
CTIC’s work will go one step further by developing a multilingual and culturally aware model, adapted to the diverse realities of Europe’s workplaces. The results will help shape practical responses in the project, including inclusive prevention measures and awareness tools.
By combining knowledge from different disciplines, and by working carefully with both language and data, this work hopes to contribute to safer and more respectful workplaces for all.
