AI runs in large data centres that consume electricity and, in some cases, water for cooling. As AI use grows, so does its environmental footprint.
Where the impact comes from
- Training: building a large model uses a lot of computing power.
- Usage: every prompt uses some energy; image and video generation use more than short text replies.
- Hardware: manufacturing chips and servers has its own footprint.
The other side
AI is also used to improve energy grids, design more efficient materials, model climate and reduce waste. Providers are investing in efficiency and renewable energy, though progress varies.
Using AI more responsibly
- Write clear prompts to get good results in fewer attempts.
- Use smaller, faster models for simple tasks.
- Avoid generating dozens of images or videos you don't need.
- Reuse good outputs and prompt templates.
Keep perspective
An individual's chatbot use is small compared to many daily activities, but collective habits and industry choices matter. Supporting transparency from providers helps.