Artificial intelligence (AI) has the potential to reduce climate-heating emissions by as much as 5.4bn tonnes annually over the next 10 years – if used strategically in sectors like food, transport, and energy, according to a study by the Grantham Research Institute and cited by E&E News.

These potential savings could surpass the additional carbon output caused by the rising energy demands of AI operations, especially those of power-intensive data centres, the research suggests.

“Artificial Intelligence can play a powerful role in supporting climate action while boosting sustainable and inclusive economic growth,” said the report.

It concluded: “The world faces an unprecedented opportunity to leverage AI as a catalyst for the net-zero transition. The five impact areas through which AI can drive emissions reductions – transforming complex systems, accelerating technology discovery, influencing behaviour, modelling climate interventions and enhancing resilience – provide a clear roadmap for harnessing AI’s potential.”

AI development has expanded rapidly in recent years. President Donald Trump has stressed its strategic value in competing with China, but its growing electricity consumption has raised red flags about the technology’s environmental cost, noted E&E. This concern is particularly significant as Trump and other figures have used AI competition as a rationale to extend fossil fuel use.

Projections from the International Energy Agency show that data centres – which are themselves proliferating – may double their power use by 2030. This surge in demand is already putting pressure on electricity systems in the US. For example, oil firms are deploying AI to identify new drilling sites.

BloombergNEF has reported that most of the additional electricity needed for data centres in the next decade is expected to come from fossil fuels – making carbon reduction goals harder to meet, said E&E.

However, the Grantham study points to ways of minimising AI’s environmental footprint.

It identifies five areas where AI could be used to reduce emissions – spanning behavioural change, innovation and energy system improvements.

For instance, smarter forecasting enabled by AI could help electricity networks better predict demand and generation, particularly from renewable sources like wind and solar. That would help utilities rely less on backup fossil fuel plants, smoothing energy flows and lowering overall emissions.

“Power grids are at the heart of the entire economy, so improving their efficiency reduces emissions across multiple sectors,” Roberta Pierfederici, a policy fellow at the Grantham Research Institute and the study’s lead author, told E&E via email.

In agriculture and food systems, AI could help develop plant-based proteins to replace animal products, which are responsible for substantial greenhouse gas output. The technology might also aid transport decarbonisation, either through enhancements to electric vehicle batteries or by encouraging shared mobility options.

All of these measures combined could curb carbon output by 3.2 to 5.4bn tonnes a year by 2035 – exceeding the total annual emissions of the European Union. For comparison, US emissions in 2023 were 6.2bn tonnes.

Though these reductions would not be enough on their own to meet the global target of limiting warming to 1.5°C since preindustrial times, they could counterbalance the emissions produced by AI itself. The study estimates that emissions linked to AI and data centres could range between 0.4 and 1.6bn tonnes over the next 10 years.

The authors acknowledge limitations in their analysis, noting that their estimates only account for AI’s impact in three sectors and don’t factor in the possibility that gains in energy efficiency could cause greater demand and emissions elsewhere. At the same time, they said they might have underestimated how broadly AI could contribute to lowering emissions as the technology evolves.

Pierfederici noted to E&E that the rising emissions from digital infrastructure are indeed a valid concern. Still, she believes the findings show AI can be part of the solution: “That said, governments need to play an active role in guiding how AI is applied and governed, to make sure the downsides are managed effectively and the full potential of AI for climate action is realised.”