Sustainability
What is it?
Sustainability is one of the most overlooked yet crucial challenges in the use of modern artificial intelligence. AI requires enormous computing power—both when models are trained and when they are used. This leads to high energy consumption, a significant CO₂ footprint, and often large amounts of water used to cool servers. This is especially true for large language models and image generators that are widely used today.
Artificial intelligence is not an “invisible” or “clean technology”—it runs on servers in massive data centers that require electricity. If that electricity doesn't come from renewable sources, it has a direct negative impact on the climate. In addition, the production of specialized hardware (like GPUs) requires the extraction of rare metals, contributing to environmental impacts through mining and manufacturing.
Examples:
Training a large language model like GPT-4 is estimated to have used as much electricity as several thousand households consume in a year. On top of that comes the ongoing power consumption needed to keep the model running and answering user queries around the clock worldwide.
In drought-affected areas—such as certain U.S. states or parts of South America—there has been criticism of the large amounts of water used to cool data centers. Water that could otherwise be used for agriculture or drinking.
Even in smaller, more localized applications—like companies generating thousands of marketing images or automating large volumes of text—the total energy consumption can be high. That's why it's important to consider how and when AI is used.
What to consider?
Ask yourself whether your AI project truly requires large models, or whether lighter and more energy-efficient alternatives exist. Open source models, edge computing, or traditional automation might be sufficient.
Check your vendors: Do they use renewable energy? Do they have a green strategy? Also consider bundling tasks and reducing unnecessary use—especially during experimentation and testing.
If sustainability is part of your company’s core values, you should integrate it into your AI strategy. This applies to your choice of tools, your communication, and your overall approach to responsible technology development.