I have been using AI intensely to build, code, prototype, learn, and publish. But I also had to pause and ask what this acceleration costs. Not only to my nervous system, but to the environment. AI is not floating in the cloud. It runs on electricity, water, minerals, servers, cooling systems, land, and labour.
The IEA estimates data centres used around 415 TWh of electricity in 2024, about 1.5% of global electricity use, and says demand has been growing fast. It projects electricity generation for data centres could rise to over 1,000 TWh by 2030. MIT also notes that generative AI has environmental consequences through electricity demand and water consumption, especially because inference happens every time people use the model, not only when models are trained.
AI helped me code, prototype, design, learn, and build faster than I ever could before. It opened doors for someone like me, a mother, immigrant, neurodivergent designer, working with limited time and big ideas.
But I also needed to pause because I am exhausted and thinking about what the future holds for my daughter.
AI is not weightless. It is not just ‘the cloud’. It depends on data centres, electricity, water, minerals, cooling systems, and infrastructure most of us never see. The same tools that help us create can also extract from the planet when used without limits.
So I am trying to move from compulsive AI use to conscious AI use.
Less prompting for noise, moving back to Procreate, Dovetail, Figma without Figma Make. More prompting for work that matters, so I am limiting the days..
Less endless generating. More editing, nurturing, learning, and publishing what already exists.
Less chasing speed. More asking whether this output deserves the resources behind it.
I still believe AI can help people build, especially people who were locked out of traditional systems
Read more:
https://news.mit.edu/2025/explained-generative-ai-environmental-impact-0117?utm_source=chatgpt.com