AI's energy and water consumption is accelerating faster than climate efforts can offset it, and every online query you make contributes to the problem. Data centers powering artificial intelligence used 448 trillion watt-hours of electricity last year, more than all but 10 countries in the world, according to a new report from the United Nations University. By 2030, data centers alone will require nearly 2.5 trillion gallons of water just for cooling, equivalent to drinking water for the entire world for 1.7 years.

The scale becomes personal when you consider individual usage. Getting a single AI text response from ChatGPT consumes about as much energy as running an efficient light bulb for two and a half minutes. That might sound trivial until you realize ChatGPT alone processes 2.5 billion queries per day. Generating a complex video through AI, by contrast, requires the equivalent of 42 hours of that same light bulb running and uses a gallon of water.

Why Are AI Companies Keeping Energy Use Secret?

One major obstacle to addressing the problem is transparency. Private AI companies rarely disclose how much energy or water their services consume. Researchers attempting to calculate these costs have found almost no public information beyond occasional blog posts. This opacity forces experts to make estimates based on open-source AI models, which may not reflect the true environmental impact of commercial systems.

"We have no way of knowing and getting a sense of the amount of energy," said Mosharaf Chowdhury, a computer science professor at the University of Michigan who tracks energy consumption of open-source models.

Mosharaf Chowdhury, Computer Science Professor, University of Michigan

The lack of transparency extends to how AI is integrated into everyday tools. Many search engines, including Google, now automatically route queries through AI without users requesting it, making it difficult for people to understand the environmental cost of their searches. Users must actively opt out rather than opt in, shifting the burden of choice away from individuals.

What Practical Steps Can You Take to Reduce Your AI Footprint?

Despite the scale of the problem, experts emphasize that individuals are not powerless. Several concrete strategies can reduce your environmental impact without requiring you to abandon AI entirely.

• Use AI less frequently: The simplest approach is to avoid AI for tasks you can accomplish through traditional methods. Don't use AI for calculations, directions, store hours, recipes, or shopping lists, which people historically managed without machine learning.
• Keep queries concise: When you do use AI, provide only essential information. Unnecessary background details force the system to process more data, consuming additional energy and water. There's no need to be polite or provide context the AI doesn't need.
• Opt out of automatic AI integration: You can disable AI in Google searches by adding "-ai" at the end of your search query or by clicking "Web" in search options. Alternative search engines like Ecosia, DuckDuckGo, and Startpage offer no-AI options or use less energy-intensive AI.

"The cleanest form of AI use is no use. So when you could avoid using AI, don't use it," said Kaveh Madani, a water scientist and director of the United Nations University Institute for Water, Environment and Health in Canada.

Kaveh Madani, Water Scientist and Director, United Nations University Institute for Water, Environment and Health

Sasha Luccioni, co-founder and chief scientific officer of the Sustainable AI Group, emphasized that many everyday uses of AI are unnecessary. "You really don't need Claude to generate a chocolate chip cookie recipe when books still exist," she noted, pointing out that productivity gains from AI are limited to a small percentage of current use cases.

Can Consumer Pressure Actually Change How Companies Operate?

While individual choices matter, experts stress that systemic change requires corporate accountability. Ana Pinheiro Privette, a former top sustainability official for Amazon Web Services, emphasized that consumers have power through market signals. "The big power I think the consumer has is the market message because I've seen that when I worked at Amazon, they listen," she explained.

Data center placement has already become a flashpoint. Years ago, companies could build data centers with minimal community pushback. Today, as facilities multiply in densely populated areas, residents are speaking up. Data centers in two Virginia counties near Washington used 2.1 billion gallons of water in 2023 alone, prompting local concern. When communities voice opposition, companies respond by implementing more efficient technologies and reducing water consumption.

"AI is not going anywhere. It has to be done. But it has to be with the help of the community, where we're understanding the concerns of the community," said Balaji Tammabattula, chief operating officer of BaRupOn, which builds energy-efficient data center campuses.

Balaji Tammabattula, Chief Operating Officer, BaRupOn

The challenge remains urgent. Data center electricity consumption is expected to more than double in the next four years. Without transparency from AI companies and deliberate choices by users, the environmental cost of artificial intelligence will continue to climb, offsetting gains made in other areas of decarbonization.

The path forward requires action at multiple levels: individuals making conscious choices about when to use AI, companies disclosing their environmental impact, and communities holding corporations accountable for resource consumption. Until then, every query carries a hidden environmental price that most users never see.