Artificial intelligence could help renewable energy systems operate more efficiently, but a new study suggests AI’s climate benefits are at risk as fossil fuel use expands. Researchers modelled 64 possible scenarios and found that AI could lead to an additional 0.47 to 1.8 gigatonnes of annual carbon emissions, equivalent to roughly 1-5% of the energy sector’s yearly emissions. The findings highlight a less-discussed side of AI’s environmental impact: its ability to make fossil fuel extraction faster, cheaper and more productive.

The research, published in Nature Sustainability, examined both the potential benefits of AI for clean electricity and its applications across the fossil fuel industry. Unlike earlier studies that largely focused on electricity consumed by data centres or potential renewable-energy savings, the researchers also considered how AI could increase productivity in drilling, exploration and extraction.

<img class="aligncenter wp-image-88432 size-full" src="https://sigmaearth.com/wp-content/uploads/2026/08/Indias-Climate-Progress-720-x-480-px-1200-x-628-px-1080-x-1080-px-Website-2026-08-13T122233.101.png" alt="AI’s Climate Benefits At Risk As Fossil Fuel Use Expands” width=”1366″ height=”768″ />

AI Could Make Fossil Fuel Production More Efficient

The researchers found that emissions declined only in scenarios where AI did not increase productivity in the fossil fuel industry. If renewable energy and fossil fuel facilities adopted AI at comparable rates, the productivity gains from clean energy would need to be at least four times greater than those achieved in fossil fuel production simply for the overall emissions impact to break even.

That creates a significant challenge because AI applications in oil and gas are already moving beyond experimentation. Companies are using the technology for seismic analysis, exploration, well planning and operational optimisation.

Oil and Gas Companies Are Already Using AI

The study’s authors argue that fossil fuel applications should not be treated as a distant possibility. Oil and gas companies are already deploying AI and advanced digital technologies to improve exploration and production.

• Saudi Aramco has said it has integrated AI across its operations and linked the technology to greater productivity and increased drilling activity.
• Norway’s Equinor has also attributed discoveries on the Norwegian continental shelf to new seismic technologies and AI, including the Lofn and Langermann wells.
• Rystad Energy estimated that digitalisation and AI could create almost $500 billion in cumulative value for fossil fuel exploration and production companies between 2026 and 2030, through increased production, more efficient operations and shorter project development timelines.
• The International Energy Agency has separately estimated that AI could increase technically recoverable oil and gas reserves by around 5% and reduce costs associated with deepwater offshore projects by approximately 10%.

Also Read: Canada Faces Severe Impact from Wildfires Affecting First Nations

Data Centres Are Only Part of the Problem

The environmental discussion around AI has largely centred on the rapidly increasing electricity demand of data centres. Many facilities rely partly on fossil-fuel-generated electricity, raising concerns about their direct emissions.

However, the researchers did not include data-centre electricity demand in their calculations. Even without it, they found that AI-enabled productivity improvements in fossil fuel production could generate emissions at least three times current estimates associated with data centres.

That distinction broadens the debate. The climate footprint of AI is not simply about how much electricity computers consume; it also depends on what economic activity those computers make more efficient.

Also Read: France Faces Severe Water Shortage as Drought Dries Up Hundreds of Rivers

Researchers Urge a Broader View of AI’s Climate Impact

The researchers described their findings as a directional and structural assessment rather than a precise forecast. Nevertheless, the relationship remained consistent across all the scenarios and sensitivity tests they examined.

The study suggests that AI’s climate benefits are at risk as fossil fuel use expands, and that the question is not simply whether algorithms can make solar panels or electricity grids more efficient. If the same technology simultaneously helps companies find more oil and gas, increase production and lower extraction costs, its overall climate impact can move in the opposite direction.

Also Read: Comparison of Smart Thermostats: Features and Costs