As AI models become cheaper and more attractive, the report predicts that this will drive new and increased usage, eroding and in some cases erasing savings from efficiency advances.
To avoid falling into this trap, we offer a roadmap for the responsible use of AI, based on the principles of transparency, efficiency by design, fairness and justice, lifecycle responsibility, global cooperation, and sustainable use.
scale of the problem
Last year, data centers already consumed as much electricity as Saudi Arabia, the world’s 11th largest electricity consumer.
If electricity use were to double as predicted by 2030, the associated carbon emissions would require growing 6.7 billion trees over 10 years to offset this demand.
Data centers would also require 9.3 trillion liters of water and nearly 10 times the land area of Mexico City.
Beyond resource use, the report also highlights the structural inequalities at the heart of the AI boom, with only 32 countries hosting AI-specific cloud infrastructure, with 90% of that capacity in the US and China.
The report warns of a growing digital divide between countries that build and control AI systems and those that use them, with the latter often bearing a disproportionate environmental burden caused by mineral extraction and e-waste.
Responsible use of AI
There are two main factors shaping the operational footprint of AI. It’s all about how much AI you use and how you use it.
This includes all tasks performed by AI models, from text and code generation to images and videos. Each of these tasks requires different levels of computational complexity.
Model selection is also important, as each AI system performs these tasks at different energy and environmental costs.
The report argues that responsible AI requires full value chain governance, from mineral sourcing to recycling and safe disposal.
This requires both the ability and environmental stewardship to think about both what AI can do for us and the protection of the natural environment.
This means making environmental information disclosure, both at the model and task level, a routine part of AI development and incorporating projected AI demands into climate and energy planning.
Responsible AI is critical as countries promote and deploy AI across governments and the public sector.
In Aotearoa New Zealand, the government has launched a National AI Strategy and a Public Service AI Framework.
The framework is based on the OECD’s values-based AI principles, including inclusive and sustainable development, but there are no environmental disclosure obligations and no regulatory authority to aggregate energy use or emissions.
In Australia, improving public services is similarly part of the national AI plan. For example, the Australian National Film and Sound Archive created Bowerbird, a machine learning-enabled large-scale audio and video transcription engine for documenting materials. The Department of Veterans Affairs has developed a proof-of-concept tool to see if AI can help speed claims processing.
Both countries have deliberately adopted a “light touch” and principles-based regulatory approach to AI. However, this approach risks overlooking the increasing environmental costs of AI that cannot be resolved by improved AI.
The natural environment is the foundation of our economy, culture, and welfare. It should be at the center of our thinking. It’s time to rethink your AI innovation strategy and shift your focus to a sustainable technology future.
Sign up to check out the discussions that dominated this week. opinion newsletter – A weekly roundup of the best commentary.
