Artificial Intelligence

The laboratory studies and develops solutions applying artificial intelligence to multiple fields of technology, with the aim of supporting the work of the Group's business areas, improving the quality and functionality of its products and services, and simplifying internal processes. The AI technologies developed in the laboratory are not only powerful enablers of products and services; they are in line with the ethical criteria of transparency and accessibility, in order to inspire confidence in the technology, promote correct interaction between humans and machines, and allow people to exploit the full potential of artificial intelligence for the benefit of all stakeholders.

Search Areas

Trustworthy artificial intelligence

We develop secure, reliable, ethical artificial intelligence systems combining robust solutions with fairness, interpretability, and respect for privacy, in order to mitigate the risks associated with implementation of artificial intelligence and increase trust among users.

Computer vision

We experiment with machine vision techniques to allow computers to understand and interpret visual information from the surrounding environment and act or formulate reports based on the data acquired.

Natural language processing

We study natural language processing to enable machines to understand, generate and interact with human language. Our research spans areas such as language understanding, generation, translation, and analysis of sentiments and dialogue systems, for application to virtual assistants, chatbots and language translation.

Multimodal learning

We explore multimodal learning techniques to allow artificial intelligence systems to effectively process and interpret different sources of information, such as texts, images, and audio, in order to obtain a more complete understanding and interaction of the available data.

Artificial intelligence for radar systems

We study the use of artificial intelligence techniques to improve the capabilities and performance of radar systems. In particular, we explore new algorithms for target anomaly detection, tracking, classification, and identification, enabling radar systems to operate more efficiently and adaptively in complex environments.

Reinforcement learning

We study algorithms and methodologies for training artificial intelligence software to make sequential decisions in dynamic environments, applying learning techniques for repeated trial and error interactions. The aim is to develop artificial intelligence systems capable of autonomous decision-making processes and adaptive behaviours.

Continual learning and unlearning

We develop techniques to make AI systems adaptive, reliable, and ethically sound, in order to respond appropriately to real-world stresses. We work both with continuous learning, for adding new knowledge, and with selective unlearning of specific segments of knowledge, such as prejudices and incorrect or obsolete information.

Artificial intelligence for cybersecurity

We develop artificial intelligence-based solutions to detect, prevent and mitigate cyber threats, with the aim of improving the resilience and security of digital systems and networks. Our research covers all phases of cyber defence, from anomaly detection to threat intelligence and automated response systems.

Roberto Cingolani, at “Frame 2025 - Being Human in the Age of Artificial Intelligence”

Leonardo’s HSE management systems certification has been confirmed

“The science of vertical flight”, Leonardo’s new educational initiative

Leonardo leads the helicopter supply for energy industry support

Leonardo signs MoU in the Kingdom of Saudi Arabia to expand collaboration in the aerospace and defence sector

Italy’s Mattei Plan: Leonardo delegation in Ivory Coast

Artificial Intelligence

Search Areas