AI-Driven Fossil Gains Raise Emissions

A peer-reviewed study finds that artificial intelligence could increase global energy-related emissions if productivity gains in oil, gas, refining, and power generation keep pace with or exceed gains in renewable energy.

Across 64 modeled scenarios, fossil-fuel expansion produced estimated annual emissions increases of roughly 0.5 to 1.8 gigatonnes, while renewable improvements would need to advance several times faster to offset them. The findings shift attention beyond data-center electricity use toward emissions enabled by AI applications in fossil-fuel production, including questions about disclosure, accountability, and limits on such tools.