Study Finds AI Productivity Gains in Fossil Fuel Sector May Outweigh Climate Benefits

A new study published in the journal npj Climate Action finds that AI driven productivity gains in the fossil fuel industry could increase global energy emissions more than AI helps reduce them through clean energy improvements. Researchers modeled 64 scenarios and found that net yearly carbon pollution could rise by 0.47 to 1.8 gigatonnes, which represents approximately 1 to 5 percent of the energy sector's annual emissions. This increase is attributed to AI enabling the oil and gas industry to produce more fuel more efficiently.

The research was conducted by authors including Holly and Will Alpine, former Microsoft sustainability workers who left the company to campaign against enabled emissions. They argue that while tech companies focus on the operational emissions of data centers, they often ignore the emissions enabled by the tools they provide to the fossil fuel industry. The study suggests that for emissions to break even, productivity gains for renewables would need to outpace those for fossil fuels by at least four times if adoption rates remain similar.

Industry data supports the trend of AI integration in energy. The International Energy Agency estimates AI could boost recoverable oil and gas reserves by 5 percent. Companies such as Saudi Aramco and Equinor have reported using AI for productivity and discovery. Rystad Energy estimated that digitalization and AI could create nearly 500 billion dollars in cumulative value for fossil fuel exploration and production companies between 2026 and 2030.

Both reports highlight a symbiotic relationship between the tech and energy sectors. For example, Chevron and Microsoft recently confirmed a deal to build a gas plant in Texas to power data centers, with Chevron intending to use some of that compute power to enhance its own internal AI capacities.