Artificial intelligence could increase global carbon emissions by making fossil fuel extraction faster, cheaper and more profitable, according to research challenging Big Tech’s focus on the electricity and water consumed by data centres.
Climate debate overlooks how AI is used
Public discussion about AI’s environmental impact has concentrated heavily on the rapidly expanding data centres required to train and operate advanced models.
Holly Alpine, a former senior sustainability manager at Microsoft, argues that this framing allows technology companies to ignore a potentially larger source of pollution: the emissions enabled when their AI and cloud products help oil and gas companies discover and extract additional fossil fuels.
Alpine said focusing only on operational emissions creates dangerous consequences because it measures how AI is powered without properly examining what the technology is being used to accomplish.
Research predicts higher global emissions
A peer-reviewed study published in npj Climate Action assessed AI as a productivity tool operating across both fossil fuel and renewable energy industries.
The researchers examined 64 scenarios covering different levels of AI adoption. When productivity improved across fossil fuels and renewable energy simultaneously, the model projected an annual net increase of between 470 million and 1.8 billion tonnes of carbon dioxide.
That represents approximately 1.2 to 4.8 per cent of global energy-related emissions recorded in 2024.
The study estimated that emissions enabled through fossil fuel productivity could be between 3.3 and 13.3 times greater than current emissions associated with data centres.
Oil companies accelerate exploration
Energy companies are using AI to analyse seismic information, identify potential deposits, predict equipment failures and improve drilling efficiency.
Processes that previously required months can sometimes be completed in weeks. This reduces the cost of exploration and makes deposits commercially viable that might otherwise have remained underground.
Industry estimates suggest that advanced recovery technologies, including AI-driven methods, could provide access to at least 470 billion additional barrels of oil from existing fields.
AI can also reduce methane leaks and improve operational efficiency. However, the researchers argue that lower emissions per barrel do not necessarily reduce total pollution if the same technology enables companies to produce and sell substantially more oil and gas.
Renewables face an unequal contest
Artificial intelligence can improve electricity forecasting, manage power grids and reduce downtime at wind and solar installations. Technology companies frequently present these applications as evidence that AI could accelerate the energy transition.
However, the study found a strong imbalance between the two sides. Productivity gains in renewable energy would need to be four to five times greater than those achieved in fossil fuels before the overall emissions effect reached break-even.
This reflects an economy in which fossil fuels still provide more than 80 per cent of primary energy and remain deeply embedded in transport, industry, chemicals and manufacturing.
Big Tech accounting questioned
Holly and Will Alpine left Microsoft after campaigning internally for greater scrutiny of the company’s relationships with fossil fuel producers. They later established the Enabled Emissions Campaign.
The organisation argues that technology companies should disclose pollution resulting from the use of their products to increase fossil fuel production. These “enabled emissions” are not generally included in corporate climate accounts, which concentrate on direct operations, purchased electricity and supply chains.
The research is based on economic modelling rather than a precise forecast, and real-world results will depend on regulation, energy prices and the speed of renewable deployment. Some energy analysts also argue that AI could accelerate low-carbon technologies more rapidly than the model assumes.
Calls for stronger governance
The researchers recommend that governments classify AI applications supporting fossil fuel expansion as high-risk and require technology companies to assess their broader climate consequences.
They also argue that improving renewable technology alone will be insufficient without restrictions on AI-enabled fossil fuel production.
The central warning is that AI does not automatically support or damage the climate. It strengthens the economic system in which it is deployed. Without deliberate policy controls, that system remains overwhelmingly dependent on fossil fuels.
Newshub Editorial in North America – 19 August 2026
Ask NF GPT
If you have an account with ChatGPT you get deeper explanations,
background and context related to what you are reading.
Recent Comments