AI‑driven productivity gains in the oil and gas industry will largely increase global emissions, according to a peer‑reviewed report led by former Microsoft employees and published in Nature. Upstream extractors can use AI to cut drilling costs, find new deposits and improve recovery rates, making fossil fuels cheaper and more abundant. While AI can help solar and wind operators forecast supply, do predictive maintenance and optimise battery dispatch, the report finds these savings are outweighed by a larger expansion of fossil‑fuel supply.

If AI adoption spreads across all energy sectors, it could add between 0.5bn and 1.8bn tonnes of CO₂ to the atmosphere each year, the authors say, compared with a potential avoidance of up to 500 million tonnes from renewables. Each 1 % productivity gain in fossil fuels would require a 4‑5 % gain in renewables to keep net emissions neutral. The study notes that the emissions impact of energy companies using AI is 2.8 to 10 times higher than prior estimates for data‑centre emissions.

Consultancies such as Wood Mackenzie project that AI could unlock 470 bn‑1 tn extra barrels of oil, while Goldman Sachs estimates an 8‑20 % reserve expansion and a 30 % cost reduction in shale. Only 44 % of upstream firms currently use AI, with 45 % planning to adopt it within three years.

The report calls for climate‑AI policy frameworks that consider "enabled emissions" and combine carbon pricing with limits on AI‑enabled fossil‑fuel expansion.