When we think of artificial intelligence, we often picture cutting-edge innovation, jaw-dropping AI-generated visuals, or virtual assistants that seem to understand us better each day. But have you ever paused to consider the environmental toll of all this digital wizardry? Here’s a surprising fact: generating just a single AI image can use up to half a litre of water. Yes, really.

At Jelliby, we believe in smart technology — but also in responsible technology. That’s why today, we’d like to shed light on a topic that’s becoming increasingly critical to the future of digital development: AI sustainability.

Why Does AI Need Water to Work?

Most people don’t associate the digital world with the consumption of physical resources — but the reality tells a different story. Every time an AI generates an image, it relies on vast servers and data centres that must be kept at optimal temperatures to run efficiently. And to regulate those temperatures? You guessed it: water. A lot of it. 

According to Shaolei Ren, a professor of engineering at the University of California, Riverside, generating between 10 and 50 prompts with GPT-3 can use roughly 500 millilitres of fresh water. Given that GPT-4 is significantly more powerful, it’s fair to assume that figure only increases.

In other words, producing just a few AI-generated images or text responses can consume the equivalent of a glass of drinking water. Now, multiply that by millions of users interacting with AI systems every single day.

Tech That’s Aware of Its Environmental Footprint

In a world where the climate crisis is very much at our doorstep, sustainable technology is no longer a nice-to-have — it’s a must. At Jelliby, we think it’s time to start asking the right questions:

  • How are the data centres we use being run?
  • What AI models are we choosing for our work?
  • Can we streamline our digital processes to lessen their environmental impact?

The answer is yes. More sustainable options do exist — from zero-emission cloud providers to prompt and algorithm optimisations that require fewer computing resources.

Artificial intelligence is developing at a rapid pace, but that doesn’t mean we can ignore the environmental cost. At Jelliby, we take a conscious approach to technological efficiency and environmental responsibility.

Can We Innovate Without Leaving a Heavy Water Footprint?

In short: yes — but it requires both awareness and action. In an increasingly data-driven world, responsible innovation must go hand-in-hand with environmental stewardship. Technology can’t keep advancing at the expense of the planet’s natural resources.

At Jelliby, we’re committed to championing AI that is efficient, transparent, and environmentally conscious — because truly smart technology should also be sustainable.