AI's carbon footprint could equal 10 million cars on the road by 2030, study warns

A study published in Nature Sustainability warns that the carbon footprint of artificial intelligence (AI) could equal that of adding 10 million cars to U.S. roads by 2030. The rapid expansion of AI computing, fueled by advancements such as ChatGPT, has created a substantial demand for energy-intensive data centers, resulting in significant carbon emissions and water consumption.

The research highlights the need for strategic solutions to mitigate the environmental impact of AI deployment. Relocating data centers to regions with abundant renewable energy potential and minimal water scarcity, such as the Midwestern "windbelt" states, could significantly reduce carbon emissions and water usage. However, even with efficiency improvements and renewable energy investments, the AI industry faces challenges in meeting sustainability targets and preventing a rebound effect that increases overall computing demand and carbon emissions. Urgent action and coordinated efforts across multiple fronts are essential to address the environmental costs of AI technology.