AI we can all trust

Artificial intelligence adapted to critical systems

Designing AI capable of operating in critical environments means taking into account constraints that are quite different from civilian uses.

In addition to its ability to operate in environments where resources (space, connectivity, energy, etc.) are limited, this type of AI must prove its reliability. It does that by adapting to the needs of each user, to guarantee data sovereignty, and, in the face of recurring risks of cyber-attacks, by demonstrating a high level of security.

Patricia Besson at Thales Research and Technology in Palaiseau — Patricia Besson in the Thales Research & Technology Laboratory © Thomas Behuret. Thales

Lasting trust

Patricia Besson is Head of Thales’s Reasoning and Analysis in Complex Systems Lab.

"Most of our products and systems already incorporate AI in one form or another. And because we operate in strategic industries such as aerospace, space and defence, we naturally set a high bar for ourselves: we have to ensure our approach to AI is based on the principles of validity, explainability, security and responsibility. That’s what we mean by trusted AI."

cortAIx

cortAIx brings together Thales’s artificial intelligence capabilities in the fields of research, sensors and systems.

It is a veritable accelerator for trusted AI, dedicated to equipping armed forces, aircraft manufacturers and airports with highly secure solutions, to provide them with more efficient data analysis and decision support, whilst taking into account specific constraints such as cybersecurity, embeddability and frugality.

Find out more See Advanced Tech in AI

800

AI experts

100

PhD students

No. 1

in AI for critical systems, with 250 patents filed

Writing 10 rules is hard enough, but writing 1,000 rules without each new rule causing a cascading effect on the others is a real mathematical challenge..

David Sadek - VP, Research, Technology & Innovation for AI & Information Processing

Research & innovation

AI we can trust … especially at decisive moments

The critical systems developed by Thales call for the very highest levels of reliability, which is why trusted AI is crucial.

Read the full interview

Tomorrow’s technology: Artificial intelligence

Explore the work of our researchers

Robust-by-design AI

Meet Teodora Petrisor, AI algorithm validation roadmap expert.

Knowledge of the real world will be incorporated into artificial neural networks to better interpret data and improve the speed and accuracy of systems in critical fields.

More on YouTube

Trustable AI validation methodology

Meet Simon Fossier, Trustable AI validation methodology expert.

AI that has proven its reliability in real-life situations will help prevent the formation of contrails through the modification of aircraft trajectories, thus making a positive impact on the environment.

More on YouTube