Publicis and AXA Measure the Ecological Impact of AI

Publicis and AXA Measure the Ecological Impact of AI

Publicis, AXA, Engie, and La Poste have launched the GenAI Footprint Alliance, joined by Accor, FDJ United, Orange, L’Oréal, and Renault Group. The goal is to create an open-source tool to estimate the environmental impact of advertising and marketing content produced with generative AI.

Generative AI promises to produce images, videos, and advertising campaigns at an unprecedented speed. However, behind every visual created in a matter of seconds, every automated spot, and every mass-generated campaign variation, one question remains largely unanswered: how much does all of this actually cost the environment?

This ambiguity is precisely what the GenAI Footprint Alliance aims to address, officially launched on June 18, 2026, at VivaTech by several major European groups. Its ambition is to provide brands, agencies, and advertisers with a concrete tool to measure the environmental footprint of generative AI, starting with an area where its use is already exploding: advertising.

A European Alliance to Clarify Advertising AI

At the origin of the project are Publicis Groupe, AXA, Engie, and the La Poste Group. Five other companies have already joined the initiative: Accor, FDJ United, Orange, L’Oréal, and Renault Group. Together, they aim to develop an open-source tool capable of estimating the environmental impact of generative AI in content production.

The issue is becoming difficult to ignore. AI tools are rapidly embedding themselves in creative chains, from writing to visual adaptations, including image generation, editing, and video. While the impact of text generation is starting to be better documented, the impact of images and especially video remains much more opaque.

GenAI Footprint thus aims to enable companies to quantify, compare, and manage these uses, rather than viewing them as mere invisible productivity gains. The scientific aspect of the project has been entrusted to the Sustainable AI Group, including researchers Sasha Luccioni, Boris Gamazaychikov, and Nidhal Jegham. Their work will particularly focus on the energy estimation of video generation, one of the most resource-intensive and least understood uses of generative AI.

The data from this research will then be integrated into EcoLogits, the tool supported by CodeCarbon, as well as e-footprint, developed by Publicis Sapient and made available within the Boavizta association. GenAI Footprint will also contribute to the Sustainable AI consortium supported by ADEME.