Play store
Generating an image with artificial intelligence requires powerful computer chips that consume significant amounts of electricity and produce large amounts of heat.
The data centres running AI systems therefore need extensive cooling. Many facilities use water-based cooling systems, while electricity generation and chip manufacturing can also require substantial amounts of water.
The exact environmental cost varies depending on the AI model, hardware, data-centre location, cooling technology and electricity source. Higher-resolution images, multiple versions and repeated edits generally require more computing power.
Researchers warn that the rapid expansion of AI could increase pressure on both energy and water resources. The International Energy Agency has reported a sharp rise in data-centre electricity demand, driven largely by AI and other digital services.
Technology companies are developing more energy-efficient chips, AI models and cooling systems to reduce the footprint. However, experts say greater transparency is needed to determine how much electricity and water individual AI services actually consume.
With billions of AI-generated images and other digital outputs being created, even relatively small resource requirements per request can add up to a significant environmental impact.