Training the future experts of trustworthy AI
Discover
Led by IRT SystemX and CentraleSupélec Exed, the Trustworthy Artificial Intelligence Advanced Master’s Program (Mastère Spécialisé® Intelligence Artificielle de Confiance) aims to train a new generation of professionals in the design, development, and deployment of trustworthy AI solutions for industry and services. This program was co-designed in close collaboration with CentraleSupélec and industrial partners, based on their expressed needs and the technological challenges identified within the Confiance.ai initiative.
The partnership between IRT SystemX and CentraleSupélec Exed combines the academic excellence of CentraleSupélec with the applied expertise of IRT SystemX in developing reliable and innovative AI solutions. This prestigious, certified program has been recognized by the Conférence des Grandes Écoles, which adds credibility and momentum to the initiative. The training combines cutting-edge theoretical knowledge with practical applications, with a strong focus on trust and ethics in AI systems.
The first cohort began in September 2024, bringing together 12 young graduates and professionals from across the fields of software, hardware, and systems engineering—many of whom are actively working on the integration of AI into industrial systems. Throughout the program, they will explore key topics such as AI ethics, transparency and security of AI-integrated systems, as well as the reliability, explainability, and robustness of algorithms. The knowledge shared by research engineers from the institute and the skills developed by the participants will be in high demand in the coming years.
After permeating our personal lives and leisure, artificial intelligence is now disrupting the business world—including companies working with systems that have high demands in terms of safety, reliability, and performance, and therefore require trustworthy AI adapted to their use cases.
Integrating AI into so-called critical systems is not just a technological challenge. Two additional hurdles must be addressed.
First is the ability to incorporate trustworthy AI into a systems engineering approach—from design through to testing and validation. This requires new methodologies and tools, which are being developed in programs like Confiance.ai, of which we are a partner.
The second challenge is skills and human resources. These must also be adapted to the specific needs of AI for critical systems. The specialized Master’s in trustworthy AI engineering, co-developed by CentraleSupélec and IRT SystemX, responds to this need. All the stakeholders of Confiance.ai, as well as those behind the AI Manifesto, and more broadly the industrial sector, welcome this initiative—it meets an urgent demand.
Marko Erman, SVP Chief Scientific Officer, Thales
Artificial intelligence combines applied mathematics, statistics, and computer science. It has existed for over 70 years and has made remarkable progress. AI has outperformed humans in games such as chess and Go, and has helped doctors by automatically analyzing medical images, enabling faster and more reliable diagnoses. It has also driven the creation of powerful language models and chatbots thanks to natural language processing. However, the most impressive and useful breakthroughs are yet to come. A major challenge is moving from proof-of-concept to industrial-grade, trustworthy products.
Hugues Talbot, Professor, Centre for Visual Computing, CentraleSupélec, Inria, Université Paris-Saclay