AI’s Green Climate Promise: Claims vs Reality
AI’s Growing Environmental Footprint.
As artificial intelligence (AI) becomes integrated into everyday life, demand for the infrastructure needed to power it is rapidly growing. Processing the vast amount of data required by AI consumes volumes of electricity in specialised data centres, fuelling a data centre building frenzy that contributes to increasing greenhouse gas emissions, water consumption and pressure on natural resources.
According to the International Energy Agency (IEA), global electricity consumption for data centres has grown by 12% every year since 2017. This is more than four times the growth rate of total electricity consumption, with the USA and China at the forefront of this rise in consumption. With electricity demand projected to be more than double by 2030, greater reliance on fossil fuels could make global climate goals harder to achieve.
Yet tech giants such as Google and Microsoft argue that AI could help tackle global climate change, with global greenhouse gas emissions potentially reduced by 36% by 2035. But how credible are these claims, and is AI’s green picture as straightforward as it seems?
Subscribe to our fortnightly newsletter
Get our top stories straight to your inbox every two weeks and stay informed about plastic pollution!
AI Businesses’ Climate Claims Under the Microscope
A report by the Norwegian analyst Ketan Joshi raises concerns about major tech companies’ claims about AI’s potential climate benefits. It distinguishes between traditional AI, which analyses data to identify patterns and make predictions, and generative AI, such as ChatGPT and Gemini, which generates original content such as images and text. Out of the 154 climate-benefit claims examined, an overwhelming 97% related to traditional AI, while only 3% concerned generative AI. Yet it is generative AI that is driving much of the expansion of AI infrastructure and its associated environmental impacts.
The lack of evidence reinforcing these claims is equally troubling. Only 26% cited published academic research, while 36% provided no supporting evidence. A further 29% relied on corporate publications, many of which lacked primary assessable evidence to validate their claims.
AI Greenwashing Playbook
Major tech companies can hide their climate-benefit claims behind a veiled wall of greenwashing. At the heart of this greenwashing is “tactical vagueness”, where two types of AI are grouped under the same umbrella, allowing companies to assert AI’s climate benefits while downplaying the environmental harm associated with generative AI.
Tech companies deploying AI must move beyond vague promises and implement robust, evidence-based sustainability measures to address its environmental impacts, with greater transparency in their climate-benefit claims.
+++++++++++++++++
For more information contact: info@scarabtrust.org.uk
Images:
Modern data centre - Image by DC Studio on Magnific (ST ref: 1428)
Energy consumption and climate-benefit diagram from“The AI Climate Hoax” Report:
ID 240620-A
Comments