Artificial Intelligence (AI), the current driving force in society's digital transition, is a heavy, resource-intensive industry. It’s not green, but it has the potential to support society's green transition. AI is at the core of the combination of the digital and the green transition – the twin transition.
Artificial Intelligence is often described as something weightless - “in the cloud” -invisible and immaterial. AI is not ethereal. It is physical. It runs on electricity, minerals, water, land, and global supply chains. If machines are going to think for us, the planet should not have to pay the price.
For years, discussions about trustworthy AI have focused on human-centred concerns: privacy, bias, transparency, and fairness. Inspired by frameworks such as the Ethics Guidelines for Trustworthy AI and the EU Artificial Intelligence Act, we have asked whether AI systems are explainable, secure, and accountable. Good, but there is a missing pillar in the debate on trustworthy AI, namely, environmental sustainability.
If AI is to be truly trustworthy, it must operate within planetary boundaries. Planet-centricity must sit alongside human-centricity in the design, development, and deployment of AI systems.
AI is not just software. It is infrastructure. It is today a heavy industry.
The Energy Hunger.
Training a single large language model can consume as much electricity as hundreds of households use in a year. But training is only part of the story. The real footprint of such models lies in inference – the billions of daily prompts, searches, translations, predictions, and recommendations that occur after deployment.
According to the International Energy Agency, data centres already account for a significant share of global electricity demand. This share is expected to rise sharply towards 2030 as AI becomes embedded across all sectors of society and across humanity. If that electricity comes from burning fossil fuels, AI risks becoming a climate-change accelerator.
The Thirst of the Machines.
Cooling data centres requires enormous amounts of water. In regions already experiencing drought and water stress, the expansion of AI infrastructure creates new pressure on ecosystems and communities. The environmental cost of AI is not only measured in its carbon footprint – it is also measured in water scarcity.
The Hunt for Critical Minerals.
The hardware for running AI systems depends on access to critical minerals extracted from a vulnerable natural environment. Extracting these minerals often causes biodiversity loss, pollution, and serious human rights challenges in global supply chains. Before an AI system ever answers a question, it may already have accumulated a significant ecological debt.
Can AI Extend the Age of Oil?
One of the most uncomfortable questions is whether AI could delay the green transition rather than accelerate it. Major oil and gas companies are among the most sophisticated adopters of AI. Machine learning helps them analyse seismic data, identify new reserves, and optimise extraction from mature wells. By lowering costs and improving efficiency, AI can make fossil fuels more competitive compared to renewables.
At the same time, the explosive growth of data centres requires a stable, 24/7 electricity supply. Without sufficient production and storage for renewables, this demand risks prolonging reliance on natural gas, or even coal, to ensure necessary baseload power.
AI can either support squeezing the last drops of carbon from the earth or help phase out fossil fuels faster. The choice is political and ethical!
The Rise of Frugal AI
The dominant trend in AI has been “bigger is better.” Larger models, more data, more computation. That trend must change.
The move from large models to smaller, specialised models – from brute-force computing to knowledge-informed computational efficiency has the potential to dramatically reduce energy use. Running AI on local devices (“edge intelligence”) rather than at large, centralised data centres will reduce both energy demand and latency.
International collaboration and open innovation, including the development of open-source software, can further prevent redundant training of similar large models, with the potential of saving enormous amounts of resources that have to be harvested from a vulnerable nature.
Smarter Hardware
Emerging technologies such as neuromorphic computing (inspired by the human brain), optical computing (using photons rather than electrons for data transport), and even quantum computing promise significant gains. While still developing, these technologies signal that AI does not have to be permanently tied to today’s energy-intensive data and computing architectures.
Circular Data Centres
We will still need data centres, but they have to be redesigned. Waste heat can support district heating systems. Backup power can rely on green hydrogen instead of diesel generators. Data centres can become integrated components of local energy ecosystems rather than isolated energy sinks.
Governance: Technology Alone Is Not Enough
Technological innovation will not automatically make AI sustainable. Governance is essential. Environmental impact must become a formal part of AI accountability. Environmental, Social and Governance (ESG) frameworks should include mandatory reporting of energy use, water consumption, and material footprint. Standardised metrics, labelling schemes and easy-to-use key performance indicators can empower users to choose greener AI providers. One bold (not new) idea is progressive taxation of digital resource consumption, keeping basic access and use affordable while discouraging wasteful overuse.
The Twin Transition
The world is digitalised, but society's digital transition continues at full speed. The twin transition is about how the dynamics and strength of the digital transition affect the green transition of society, and how these two transitions mutually influence each other and should be combined.
AI is not green, but it has the potential to be a catalyst for the green transition. AI, as of today, is therefore at the core of what the twin transition is about. AI has made the twin transition policies and strategies, as proposed in “The Twin Transition Century; The role of digital research for a successful green transition of society”, even more important.
Redefining Trustworthy AI
AI is currently a resource-hungry giant operating within fragile ecological systems. Yet it also holds enormous potential to help humanity manage complexity and accelerate sustainability.
The future requires frugal AI: systems that are smaller, smarter, more energy-efficient, and embedded into a circular economy. Trustworthy AI must no longer be defined only by its precision, fairness and transparency, but also by whether it safeguards the ecological foundations of life itself.
When machines think, the planet should not pay.
Related content:
AI strains planetary boundaries, however, it may also help us respect them. Curious to see how? Read the accompanying blog post “Can Artificial Intelligence Foster Sustainable Economic Growth?”.
Morten Dæhlen
Morten Dæhlen is Co-director of The Norwegian Centre of Trustworthy AI (TRUST) and professor in computational mathematics/informatics at the University of Oslo. Previously he was Centre Leader for dScience – Centre for Computational and Data Science (2021-2025), Dean at The Faculty of Mathematics and Natural sciences (2013-2020), Head of the Department of Informatics (2005-2012) and Director at SINTEF, the Research Council of Norway and Simula Research Laboratory AS.
His main research topics are computational methods - machine learning, visualization and geometric modelling. He has worked on basic scientific problems as well as applied challenges in collaboration with companies and units in public sector.