The first face-to-face workshop of the Artificial Intelligence, Scalability & Transferability (AIST) project took place on 22–23 September 2025 at the headquarters of the International Union of Railways (UIC) in Paris. This marks the launch of a three-year initiative (2025–2027) aimed at advancing the use of artificial intelligence in rail operations.
Participants included:
- Funding members: German Federal Railways (DB) InfraGo, Swiss Federal Railways (SBB), Austrian Federal Railways Infrastructure (ÖBB INFRA), Spanish Administrator of Railway Infrastructures (ADIF), Italian Railway Network (FS-RFI), Iarnród Éireann (Irish Rail),and Infraestruturas de Portugal (IP).
- Guest organisations: National Railways of France (SNCF) Réseau, British Network Rail
Key topics discussed by AIST members included:
- Scalability and industrialisation, including the Machine Learning Operations (MLOps) lifecycle
- Sharing of datasets
- Explainable Artificial Intelligence (XAI), with a focus on trustworthy AI and its acceptance by railway experts
- Lessons learned from both successful and unsuccessful projects
- Identification of quick wins
- Best practices from other sectors and industries
- Sharing of pretrained artificial intelligence models
- Future areas of exploration, such as AI standards, model transferability, and AI governance
This collaborative effort sets the stage for meaningful progress in integrating artificial intelligence into the rail sector, with a strong emphasis on scalability, transparency, and cross-border cooperation.
