AI Could Help Fossil Fuel Companies Create More Emissions

A recent study warns that artificial intelligence could exacerbate global emissions from the fossil fuel industry, rather than help mitigate them. Published in the journal npj Climate Action, the research conducted by former Microsoft sustainability workers suggests that AI use in the oil and gas sector could significantly boost energy-related emissions.

The study estimates that AI-enabled emissions could range from Mexico's annual output to Russia's, depending on the magnitude of the effect. This increase in emissions would eclipse the benefits AI brings to clean energy technologies like solar and wind. The authors argue that the relationship between technology companies and fossil fuels creates a self-reinforcing effect, where AI's productivity gains in fossil fuel extraction and development enable additional emissions, making it harder for the world to meet its climate goals.

Will Alpine, one of the authors, emphasizes that AI's impact on fossil fuel production should not be treated independently from its role in clean energy technologies. The paper's authors, who previously worked at Microsoft, argue that tech companies often overlook the emissions generated by their AI tools, referring to these as "enabled emissions."

To illustrate the potential impact, the researchers used an economic model that accounted for AI's productivity-enhancing effects across various fossil fuel industry sectors. Their findings suggest that AI could contribute between 1.2% and 4.8% to global energy-related emissions, a figure that surpasses projections for data center energy use.

Chevron and Microsoft's recent partnership, which involves building a gas plant to power data centers, exemplifies the interconnection between AI and fossil fuels. This alliance highlights the potential for AI to both advance renewable energy and increase fossil fuel production, a dynamic that the researchers caution could have detrimental effects on the environment.