The European Trustworthy AI Association Gathers Momentum
On the occasion of the Trustworthy AI Summit, the European Trustworthy AI Association is following the trajectory it initially set to develop and animate the European community for responsible and industrial AI. Its goal is to facilitate the design and validation of reliable, explainable AI systems that comply with evolving regulations, using open-source tools. The association has already announced the membership of around thirty organizations across Europe and worldwide.
AI is considered as one of the main “game changers” in the global economy, generating high expectations and numerous promises. Yet, the reality shows that the path to industrial deployment is still long. Trust is the key element in the AI value chain, and it can only be guaranteed through certified components and a methodology compliant with regulatory requirements.
This is precisely the ambition of the European Trustworthy AI Association, born from the initiative of actors in the Confiance.ai program: to offer a reference methodology and tools to promote the adoption of responsible AI and enable industrial-scale deployment.
Open Source and European Openness: The Association’s Founding Principles
The Confiance.ai program has positioned France at the forefront of trustworthy AI engineering. To maintain long-term leadership, openness is key. The association therefore positions itself within a European dynamic and aims to expand its community to a growing number of international partners. Scaling up is essential for the association’s success to ensure the quality of its tools, scientific relevance, and ability to meet industrial challenges at large scale.
Building a Large European Ecosystem Around Trustworthy AI
In this perspective, the association has launched a strong development momentum, with interest expressed by many key European actors in responsible and industrial AI. Around thirty industrial companies, start-ups, academic institutions, associations, and research organizations at the European level have already joined or are in the process of joining (Air Liquide, Airbus, Bureau Veritas, CEA-List, ECE, Eyesnap, eOdyn, KDNS, LGM, MBDA, Multitel / TRAIL, National Technology University of Athens, Naval Group, Octopize, Safenai, IRT Saint-Exupéry, Safran, Simula, Sopra Steria, Thales, University of Southampton / RAI UK, etc.), along with two international partners (the Japanese association JAMBE and the Canadian research organization CRIM). The association has also begun initial collaborations to accelerate its European development with leading European AI associations such as BDVA, Data Robotics Association, and the AI-on-Demand platform.
Open Source as a Lever
To embrace this principle of openness, the association has decided to build its strategy around open source. This approach enables:
- Collaboration and knowledge sharing
- Transparency and security in the development and validation of AI-based systems
- Greater independence and control compared to proprietary solutions
- Development of standards that provide international industrial recognition
During its initial phase, the association therefore transitioned all methodologies and the majority of technological assets from the Confiance.ai program to open source to promote wide dissemination and use.
Today, more than 25 open-source tools are available in the catalog, with 7 already supported for industrial use by association members.
Operational Services and an Active Community
The association is fully operational: it supports its members, organizes and unites the existing ecosystem, and continues to enrich its methodology and component catalog.
To engage its community, it has established five working groups:
- Industrial Use: to help members understand and support the adoption and deployment of open-source tools and software
- Scientific Animation: to share updates on the state of scientific knowledge and organize scientific cooperation with RTOs and other European associations
- Standardization: to coordinate and pool standardization efforts and ensure alignment of the association’s methodological and technological portfolio with regulations and standards
- Training: to coordinate the development of a training offer for engineers in partnership with engineering schools (e.g., Mastère Spécialisé IA de Confiance with CentraleSupélec) and universities
- Communication: to manage initiatives such as the Trustworthy AI Summit, the annual reference event for the European trustworthy AI community
A sixth working group is currently being created to support defense actors in addressing their specific challenges.
Membership Levels:
Interested companies and academic institutions can join via three levels:
- Engage: participation in all ecosystem events; access to the full methodological and software portfolio
- Use: adds initial support for portfolio usage and access to the 5 working groups
- Lead: includes dedicated support (SLA), participation in the roadmap of assets, and governance of the association
Next Steps
The association continues to mature its portfolio and complements it with partner solutions to equip European engineers with state-of-the-art open-source methodologies and tools. To facilitate usage, it is developing a training offer in collaboration with partner schools and universities. By working with related initiatives, it aims to promote its assets to AI stakeholders and grow its ecosystem. Simultaneously, it continues advocacy with policymakers and standardization bodies, contributing to standards that place trust at the core of future regulations. Long-term, it plans the next iterations of its methodology, including advancements in generative and agentic AI. Its ambition is to become the European reference body for industrial and responsible AI.
"For an industrial player like Thales, which deploys solutions within critical systems in sectors such as aerospace and defense, implementing trustworthy AI is essential and requires appropriate engineering tools and processes. We have been working on this topic for a long time around four pillars – validity, safety, explainability, and accountability – which we have further explored through our involvement in the Confiance.ai program. Today, we are actively pursuing efforts to industrialize trustworthy AI engineering by contributing to the establishment of the European Trustworthy AI Association."
"The European Trustworthy AI Association was initiated at the beginning of 2025 by six partners of the Confiance.ai program (Air Liquide, Naval Group, Safran, Sopra Steria, IRT SystemX, and Thales). We are delighted to see the enthusiasm and momentum building around our value proposition, with numerous expressions of interest from major European actors in the AI field. We invite all companies involved in accelerating AI adoption in industrial systems at the European scale to join us!"