EUROFEDOP members from across Europe and representatives of other workers’ organisations gathered in Bucharest (Romania) from 8 to 10 May 2025 to discuss “Fair and collaborative usage of AI for trade unions and social dialogue”. The seminar explored how AI is already transforming the world of work, and what this means for unions in terms of rights, risks and opportunities. It was organized by EUROFEDOP (Europese Federatie van het Overheidspersoneel), with the support of EZA and funded by the European Union.
The opening speeches were held by EUROFEDOP President Dr. Norbert Schnedl and Vice-President Prof. Viorel Rotila. Both underlined the importance of not allowing AI to replace the core of union work – real human interaction and collective voice – but instead to understand and shape AI for the benefit of workers.
Following the opening, Prof. Rotila gave the first presentation, emphasising that AI should not be feared but critically understood and strategically used by unions to increase influence, improve communication and support workers’ rights. He encouraged unions to build their own expertise and digital infrastructure to ensure autonomy and resilience.
Dr. Christina Colclough (Founder: The Why Not Lab, Denmark) then warned against the premature use of AI without strong internal capacity. Her call: Unions must first establish governance, legal safeguards, and clear values before adopting AI tools. In her view, without these foundations, AI can pose serious risks to workers’ rights and union credibility.
Prof. George Azzopardi, Associate Professor in Computer Science (University of Groningen NL), introduced concrete AI tools that can support union work – from predictive modelling to sentiment analysis. He presented real-world examples and focused on both the benefits and the dangers, urging unions to remain critical and ethical when adopting AI.
On the second day, Laurentiu Lungu, Head of the IT Department at FSSR Romania, shared practical use cases from the Romanian public sector. He demonstrated how AI-enhanced tools can improve services for union members, provided the tools are well-designed and developed with transparency. He stressed the importance of shared databases and European collaboration.
Professor Dr. Phoebe Moore (University of Essex) gave a thought-provoking keynote on the risks of AI-powered monitoring and profiling in workplaces. She criticised the vague term “responsible AI” and urged unions to be involved at every step – from procurement to policy – to protect workers’ privacy and agency.
In the afternoon, participants started the workshop phase in language-based groups (BCMS, Romanian, German, English). The aim was to reflect on the presentations and identify key questions and areas for further exploration.
On the third and final day, each group presented their reflections and proposals. These included calls for AI transparency, stronger regulation, better education within unions, and human decision- making as a non-negotiable principle. A joint conclusion was formed and suggestions for next steps were agreed.