AI will accelerate oil extraction more than green energy development

Researchers from Microsoft used a global computable general equilibrium model to analyze AI as a technology that boosts productivity in both fossil and renewable energy sectors. They found that parallel adoption of advanced AI algorithms could increase global emissions by about 0.47–1.8 billion tonnes of CO2 per year, equivalent to 1.2–4.8% of 2024 world energy emissions. The reason is that AI-driven efficiency gains in oil and gas extraction have a stronger economic effect than similar improvements in solar and wind power. AI is used for analyzing geological data, assessing reserves, reducing drilling risks, optimizing well operations and maintenance, thereby lowering production costs and making previously uneconomical resources viable. Wood Mackenzie estimates AI could make 470 billion to 1 trillion barrels of additional oil economically accessible, while Goldman Sachs says AI could increase US shale recoverable reserves by 8–20% and cut drilling costs by about 30%. An IBM survey shows 44% of oil and gas companies already use AI in exploration, with 45% planning to adopt it in three years. In renewable energy, AI can improve forecasting of solar activity and wind speed, detect faults, optimize grids and battery charging, potentially increasing productivity by up to 30% and preventing up to 500 million tonnes of CO2 yearly. However, studies show renewable productivity must rise 4–5% for every 1% gain in oil and gas productivity to offset emissions, and physical and administrative barriers remain. The researchers urge considering hidden emissions from AI applications in other sectors, which may be up to ten times higher than direct data center emissions.