Inside CeSIA, the newly established French center for AI safety

The recent establishment of CeSIA (Centre pour la Sécurité de l’IA) in Paris marks a significant development in European efforts to address artificial intelligence safety through education, research, and policy work.

The organization’s foundation: CeSIA is a new Paris-based center dedicated to reducing AI risks through a comprehensive approach that combines education, technical research, and advocacy work.

• The center’s mission focuses on fostering a culture of AI safety by educating and informing about both AI risks and potential solutions
• A team of 8 full-time employees, 3 freelancers, and numerous volunteers drives the organization’s initiatives
• The organization has established itself as a bridge between academic research, policy-making, and public awareness

Policy initiatives and partnerships: CeSIA has quickly positioned itself as a significant voice in AI governance discussions at both national and international levels.

• The organization has contributed to the development of the EU AI Act Code of Practice
• Strategic partnerships have been formed with the OECD for AI safety initiatives
• The team has organized influential roundtables bringing together key AI stakeholders
• CeSIA provides advisory support for French AI evaluation efforts

Research and development focus: Technical innovation and safety benchmarking form core components of CeSIA’s work.

• The center has published the BELLS benchmark system for evaluating Large Language Model safeguards
• Research efforts include exploration of “warning shot theory” for AI risks
• New approaches to safe AI design, including constructability methods, are being investigated

Educational outreach and field-building: CeSIA has developed a comprehensive educational program aimed at both academic and public audiences.

• Accredited university courses on AI safety have been established in France
• The ML4Good bootcamp program has been successfully replicated internationally
• Development is underway for the AI Safety Atlas textbook
• Regular events and MOOC contributions enhance the educational offerings

Public engagement and awareness: Strategic communication efforts have been implemented to reach broader audiences.

• The organization has published influential op-eds on AI safety topics
• Collaborations with YouTubers have helped reach millions of viewers
• Educational content has been designed to make complex AI safety concepts accessible to general audiences

Future implications: While CeSIA’s establishment represents a significant step forward for European AI safety initiatives, the organization’s effectiveness will likely depend on its ability to balance academic rigor with practical policy implementation while maintaining strong international collaborations. The center’s multi-faceted approach could serve as a model for similar institutions worldwide.