AI and the Environment - International Standards for AI and the Environment 2024 Report
Reference Type:
Report
This report addresses the intersection of Artificial Intelligence (AI) and environmental
sustainability, emphasizing the importance of international standards in guiding the ICT industry
towards an environmentally sustainable future. As AI continues to revolutionize various sectors
with its transformative capabilities, it also presents significant environmental challenges. These
include high energy consumption in data centres and increased greenhouse gas emissions
due to the substantial computational power required for AI operations, as well as the growing
problem of e-waste from rapidly advancing technology.
Understanding the lifecycle of AI is crucial when assessing its environmental impacts, as it
highlights stages where energy consumption and GHG emissions can be most significant
and where reductions can be made. The AI lifecycle includes several key stages: identifying
the problem, data collection, designing a model, model training, model evaluation, model
deployment, and inference. During these stages, particularly model training and inference,
substantial computational power is required, leading to high energy consumption in data
centres. This process contributes significantly to increased greenhouse gas emissions. By
comprehensively understanding these lifecycle stages, it is possible to pinpoint where energy
consumption and GHG emissions are most pronounced, enabling targeted actions to mitigate
their environmental impact and promote sustainable AI practices.
Continuous innovation in AI and its applications is crucial for tackling global challenges, including
climate change. However, without proper guidelines and standards, the environmental footprint
of AI could outweigh their benefits. This report highlights the role of international standards in
ensuring that AI developments are effective and sustainable.
The ICT industry stands at the forefront of this transformation. By adopting and adhering to
these standards, industry players can not only enhance the environmental efficiency of their
AI systems but also leverage AI as a powerful tool for climate action. Standards provide the
necessary framework to measure, manage, and mitigate the environmental impacts of AI, from
the product level to the network level, encompassing data centres and addressing e-waste
management. The ITU-T SG5 has been producing backbone standards for the environmental
efficiency of AI and other emerging technologies, ensuring a comprehensive approach to
sustainability in the sector.
This document serves as an essential resource for governments and the industry and, offering
comprehensive insights into the current standards landscape and ongoing efforts to improve
the environmental sustainability of AI. It not only details how to build and implement sustainable
practices but also emphasizes the importance of measuring progress. Accurate measurement is
crucial for assessing the effectiveness of sustainability initiatives and making informed decisions
to drive further improvements. The report also underscores the importance of collaboration
among experts, academia, member states, and standards development organizations (SDOs)
to achieve these goals. By adopting a unified approach to standardization, we can ensure that
AI contribute positively to our collective goal of achieving global sustainability targets.
Download Reference:
Search for the Publication In:
Formatted Reference: