AI models like ChatGPT have a significant and growing environmental footprint. The environmental cost of using AI tools is vast and rapidly increasing. While a single query may seem trivial, at a global scale, the weekly and annual impact is equivalent to the energy and water use of entire nations, and our AI  carbon footprint is growing.

Hello and welcome back to the AI SIX podcast, helping you understand AI in just six minutes a day. I do the hard work so you don’t have to. 

So we have an AI sustainability dilemma. 

Using ChatGPT for a week (at a global scale) is equivalent to:

• Powering a major skyscraper for months.
• Charging millions of smartphones or hundreds of thousands of electric vehicles.
• Consuming more electricity than many entire countries in a week.

The environmental cost is not just electricity: Water use for cooling and the carbon footprint from the energy mix are substantial, and both are expected to rise as AI adoption grows and models become more complex.

What are the big AI players saying and doing about this?

Meta, OpenAI, and Google acknowledge the environmental impact of LLMs and are investing in efficiency, renewables, and transparency. However, emissions are still increasing due to rapid AI adoption, and there’s a need for clearer reporting, stronger commitments, and continued innovation in green technology.

So what about our personal use? I wanted to know what my carbon footprint might be by using ChatGPT for a week. Here’s what I found. 

Using ChatGPT as an individual has a very small impact, but it’s not zero.

• Electricity: 15 prompts in a week use about the same energy as a small LED bulb left on for a few hours.
  
• Water: Your usage would require 7–8 litres of water, roughly one short shower. This water cools the data centres.
  
• Carbon: The emissions are less than driving a car for 1 kilometre.

In short, our weekly use is tiny on its own. But with millions of users, it adds up and becomes a real environmental issue globally.

What can we do to reduce our impact with AI?

How to Use AI More Sustainably (As an Individual)

• Be intentional: Use AI when it genuinely adds value, not just out of habit.
• Pick lighter models – Choose GPT-3.5 or other efficient tools for simple tasks.

• Support green platforms – Use services powered by renewable energy.
• Write better prompts – Clear, concise prompts reduce energy use.

The big responsibility is to balance innovation with sustainability.

Smarter AI models often use more energy. Companies are starting to recognise this and are working to make AI more efficient and eco-friendly.

By being aware of the impact, we can all make better choices, like using smaller models for simple tasks or writing smarter prompts to reduce energy use.