Artificial Intelligence (AI) Serving the Ecological Transition

Artificial intelligence allows for predicting and classifying to better understand the dynamics of the territory: mobility of people, land use, or distribution of species, for example. It also optimizes our processes, to reduce our resource consumption (water, energy, miles traveled) and to provide a better public service (assistance in reading environmental authorization files, for example). AI-based solutions therefore represent an opportunity for national and local public actors to address priority ecological transition challenges. It is essential to use it as a tool in response to a professional need when necessary, rather than making it a goal.

The Ministry of Ecological Transition, Biodiversity, Forests, the Sea and Fisheries, and the Ministry of Spatial Planning and Decentralization have a roadmap “AI and ecological transition” led by the Ecolab of the General Commissariat for Sustainable Development (CGDD). This roadmap is part of the National Strategy for AI, aimed at development and innovation in priority areas such as AI in the service of ecological transition.

To accelerate ecological transition in the territories, the Ecolab launched in May 2022 a “Community of AI Actors in the Territories (CAIAT) for ecological transition“, made up of representatives from territories, competitiveness clusters, research institutions, micro-enterprises/startups/SMEs, and companies. The goal of the community is to share feedback and best practices. This Community notably includes pilot territories, winners of the France 2030 project calls “Demonstrators of Frugal AI in the Service of Ecological Transition of Territories” (DIAT), which have developed experiments in real conditions on a representative scale and emphasized the reproducibility of projects.

To build the capacities of administrations, the Ecolab runs an “AI Club” bringing together general management and operators from the ministerial sector to share ongoing experiments and feedback, and potentially foster collaborations.

The Ecolab has published the first General Reference for Frugal AI, in collaboration with Afnor, to measure and reduce the environmental impact of AI.

The community of AI actors

The Ecolab aims to bring together the community of AI actors in the territories during various workshops organized each year, either in person or via webinars.

The goal of the community is to share best practices and feedback from AI projects serving ecological transition in the territories. It relies on a partnership between the Ecolab and the Interconnectés and Hub France AI to increase visibility among local authorities and the economic ecosystem.

Frugal AI

Frugal AI is a concept that encompasses the fact that an artificial intelligence solution uses fewer resources (raw materials, water, electricity…) throughout its life cycle.

According to the study ARCEP & ADEME 2023, digital technology accounts for 2.5% of France’s carbon footprint. The ecological footprint of the algorithms necessary for the functioning of AI systems is established and significant given that they are major consumers of energy, resources (water, metals, and minerals), and infrastructures (networks, data centers, and terminals). This is why Ecolab has developed a general reference for frugal AI, in collaboration with Afnor, which provides organizations wishing to develop an AI solution with a methodology for assessing environmental impact and best practices to adopt.