Every online query sent to an AI system increases its environmental footprint, as the data centers that process the requests consume rising amounts of electricity and water, researchers said. The resources needed to power and cool the facilities are straining climate and water conservation goals.
“AI is going in the opposite direction to decarbonization efforts,” said Sasha Luccioni, a cognitive computer scientist and co-founder and chief scientific officer of the Sustainable AI Group. “We should be thinking about where we are going towards. If you’re recycling and a vegan but then you’re using ChatGPT to do your multiplication for you, well that’s kind of against the trend.”
The rapid deployment of AI services has made the technology a growing contributor to the energy and water footprint of the technology sector. Data centers that train and run AI models require vast amounts of electricity for computing and water for cooling. At the same time, the companies operating these facilities have offered limited public disclosure about the scale of their resource use, experts said.
“AI companies aren’t transparent about how much of those resources they use,” Luccioni said.
Researchers said the rising resource consumption creates a tension with the push for a low-carbon economy. While some technology companies have pledged to achieve net-zero emissions or to offset their water use, the growth in AI services is making those commitments harder to meet.
The findings add to a growing body of research on the environmental cost of AI. A United Nations report released earlier this month found that the electricity consumption of AI data centers rivals that of entire countries. Other studies have documented that the majority of planned U.S. data centers are located in drought-prone regions, and that nearly 80% of the world’s data centers are exposed to climate hazards such as heat and water stress.
Luccioni said individual users also play a role. Using an AI chatbot for tasks that could be done with a simple search or calculator adds unnecessary resource demand. Reducing casual AI use, she said, is one modest step consumers can take alongside broader pressure on companies to operate more sustainably.