Unraveling the Hidden Environmental Impacts of AI Solutions for Environment

TL;DR

The paper argues that evaluating AI's environmental impact requires moving beyond energy and GHG metrics to a full life cycle assessment of AI services. It adapts LCA to AI by defining first-order impacts from hardware and use-phase dynamics, and it introduces a framework for comparing reference and AI-enhanced applications to quantify environmental benefits and costs. It critiques current AI-for-Green work for underestimating impacts, and it emphasizes data gaps and the need to consider second- and third-order effects, especially under large deployment. The work aims to establish a rigorous, multi-criteria approach that can guide the environmental valuation of AI-enabled environmental solutions, while acknowledging its limitations and the potential for socio-technical rebound effects.

Abstract

In the past ten years, artificial intelligence has encountered such dramatic progress that it is now seen as a tool of choice to solve environmental issues and in the first place greenhouse gas emissions (GHG). At the same time the deep learning community began to realize that training models with more and more parameters requires a lot of energy and as a consequence GHG emissions. To our knowledge, questioning the complete net environmental impacts of AI solutions for the environment (AI for Green), and not only GHG, has never been addressed directly. In this article, we propose to study the possible negative impacts of AI for Green. First, we review the different types of AI impacts, then we present the different methodologies used to assess those impacts, and show how to apply life cycle assessment to AI services. Finally, we discuss how to assess the environmental usefulness of a general AI service, and point out the limitations of existing work in AI for Green.

Figures (4)

• Figure 1: LCA dimensions: the first dimension corresponds to the phases of life cycle, the second one to the environmental impacts (see \ref{['sec:estimate']} for more details on this last dimension).
• Figure 2: Diagram representing the Life Cycle Inventory of an AI service: Above: an AI for green application corresponds to the inference step that depends on other unit processes that require various devices. Below: the use of devices is located in a more global environment, including production of resources and impacts. In both schemes colored boxes correspond to unit processes, black arrows correspond to economic flows (bold: material, dashed: energy) and red arrows to environmental flows.
• Figure 3: Overview of AI's impacts. First-order or direct impacts result from the equipment life cycle. Second-order impacts are the difference between the LCAs of the reference system and the AI-enhanced system. Third-order impacts are changes in technology or society induced by the application.
• Figure 4: Sankey diagram of parts of Rolnick's paper references in terms of environmental evaluation (created with the Sankey Diagram Generator by Dénes Csala, based on the Sankey plugin for D3 by Mike Bostock; https://sankey.csaladen.es; 2014)