AI is present in many systems that require interaction with humans. For this reason AI is becoming an acting element and therefore a component of multi-agent crews, with strong interdependence and mutual knowledge of the members’ operating methods, which is expected to improve efficiency and performance. In such framework, a human operator will rely on an AI system if it is certain that the provided information is reliable. Conversely if the AI system is not accepted, there is a risk of humans bypassing the system. Furthermore, human over-confidence towards AI can also endanger a crew.
Objectives
Chair NAIADE’s objective is to provide crew members with efficient means to bound their cognitive charge and allow them to focus on decision-making in a complex naval environment. The ambition of envisioned research is to improve the performance of the crew member-AI team by offering solutions for modeling interactions between crew members and AI technical tools for decision-making. In this context, the AI system is considered as an active agent interacting with a human or other AI system.
Role of the school IMT Atlantique
As part of its research activities IMT Atlantique performs algorithmic and methodological studies to facilitate decision-making, analyzed through the triptych Data - Information - Decision. This action is part of a process to reinforce research activities in the fields of AI, decision-making, and data science, examining their expansion to the maritime domain field.
Expected outcomes
To study AI systems that perform tasks initially assigned to a human operator, the applied approach will include the definition of: AI assessment models, adaptive decision-making assistance, explainability qualification and quantification models, characterization of an operator’s mental load, and adaptation of AI to facilitate the operator’s task. In addition, to analyze the AI system as an agent of the resulting socio-technical system, research will study the socio-technical system formed by humans and AI types of agents, in particular: characterize the impact of human-AI interactions on decision management, define adapted human-machine interfaces, measure the effectiveness of the various mixed teams configurations formed with AI, integrate into AI the possibility of identifying the operator’s cognitive load, and predict human-AI performance.
Partners
As part of its research activities, the Ecole navale is carrying out work on the usages’ transformations that AI driven systems are conveying to the French navy, both with regard to the organization, the nature of activities, the human/system interactions, and new developed skills. This work links design and use of systems driven by AI.
Naval Group : As an industrial prime contractor, designer and integrator for whole warships and combat systems, Naval Group is an international naval defence player and a partner for countries seeking to maintain control of their maritime sovereignty. Thanks to its extraordinary know-how and industrial resources, it designs, produces, integrates, supports, dismantles and deconstructs submarines and surface ships.
Thales : Major actor in the aeronautical, space, defense, and security sectors, seeks to develop skills on the application of AI-based systems to the maritime domain, taking into account specificities about their usage in a naval environment by crew members.
