 |
|||
|
|||
|
|||
Home > TERATEC FORUM > Workshops
 |
TERATEC 2025 Forum |
|
Defense and AI, the challenges of cybersecurity
Chaired by Patrick Fabiani, AI Roadmapping & Advanced Scientific Studies, Dassault Aviation and Omar Hammami, Chef du groupe Ingénierie Système, ENSTA Paris
Trustworthy embedded AI into safety-critical systems for aeronautics, defense and security
By Patricia Besson, Chef de Laboratoire R&T | IA | Fusion d'Informations | ModelOp, Thales
The rapid adoption of Artificial Intelligence (AI) is transforming operational capabilities in safety-critical areas like aeronautics, defense, and security. However, this change brings challenges, particularly around trustworthiness (validity, explainabilty, security and responsibility) of such AI-powered systems. Traditional verification and validation methods often fail in such a scheme, necessitating a new, rigorous approach to development, deployment, and maintenance.
To address these challenges, implementing a comprehensive AI engineering lifecycle is essential. This lifecycle should incorporate fundamental engineering principles such as safety, security, ethics, and system operational domain characterization along with specific considerations for AI algorithm development. This framework will help manage the complexities of AI technologies. Moreover, integrating MLOps/ModelOps practices within this lifecycle fosters collaboration among data scientists, engineers, and operational teams, enabling continuous integration, deployment, and monitoring of AI models.
As AI becomes more complex and embedded into critical systems, sustained collaboration among industry stakeholders, regulatory bodies, and academia will be vital. Exploring new paradigms, like runtime assurance and incremental development, underscores the need for AI systems to remain adaptable while ensuring their trustworthiness.
 |
Biography: Dr. Patricia Besson is the head of the Analysis & Reasoning in Complex Systems Lab, at cortAIx Labs France (Thales). She obtained her PhD in Multimodal Data Fusion from Ecole Polytechnique Fédérale de Lausanne (Switzerland) in June 2007. She then worked as a researcher with the Institute of Movement Sciences (CNRS & Aix-Marseille University, France) for 5 years. Her research focused on understanding and modelling the interaction between human beings and their environment through machine learning approaches, in order to develop intelligent interactive systems. From 2013 to 2017 she led a team working on data science and artificial intelligence projects (health monitoring systems, autonomous vehicles, etc.) at Safran Engineering Services (France). |
|---|
Register now and get your badge here >>>