How to mitigate the climate impact of AI

AI is changing the game in so many ways, with its potential to revolutionise industries, enhance productivity, and tackle complex problems we couldn’t solve before. While we’re excited about these possibilities, we must also consider AI's environmental impact.

AI's contribution to sustainability

While concerns are frequently raised about the negative environmental impacts of AI, it is important to recognise its significant benefits in advancing sustainability. For instance, AI can optimise resource utilisation in agriculture and manufacturing, reducing waste and improving efficiency. In renewable energy, AI can predict energy production to enhance grid management and maximise the use of wind and solar power. AI also plays a crucial role in smart cities, optimising traffic flows and enhancing public transport to reduce urban carbon footprints.

The environmental impact of AI tools

Although AI has the potential to promote sustainable practices, it's not without its environmental costs. The development, training, and deployment of AI models require substantial computational resources, which can lead to significant energy consumption and carbon emissions.

In addition to that, frequent hardware upgrades result in increased electronic waste, posing environmental challenges due to hazardous materials contaminating soil and water.

Mitigating the climate impact of AI

For AI to remain sustainable, both users and providers need to implement strategies that lessen environmental damage. Various mitigation techniques can be employed to decrease the ecological footprint of these services:

What can AI providers do?

• Develop energy-efficient algorithms

AI researchers and developers should prioritise the creation of algorithms that are more energy efficient. Specialised AI solutions using Small Language Models (SLM) can be much more efficient than Large Language Models such as ChatGPT, for example.

• Utilise green data centers

AI providers should invest in green data centers that use renewable energy sources and advanced cooling technologies to reduce their carbon footprint. Server virtualisation and dynamic resource allocation can further minimise energy consumption.

• Use new hardware specialised for AI

Utilise new chip technology designed for AI that requires significantly less energy, while also embracing circular economy principles to extend the lifecycle of hardware components through reuse, refurbishment, and recycling.

What can AI users do?

• Learn to write good prompts

Learn how to write effective prompts for AI to retrieve the information you need on the first try. Fewer interactions with the AI service contribute to greater sustainability.

• Don't use AI as a search engine

AI services require much more energy than traditional search engines. Avoid using AI tools for simple searches, like finding an address or website.

• Encourage responsible AI use

Consider the environmental impact of your AI applications by evaluating whether new model training is needed or if pre-trained models are sufficient. Also, make sure to prioritise cloud providers with strong sustainability commitments.

"Sustainability in AI is crucial. As AI advances and becomes part of more areas of our lives, it's important to recognise both its potential to support sustainable practices and its environmental impact. By focusing on sustainability, we can use AI’s power to create a future that’s not only innovative but also environmentally responsible."

About the author: Stefan Agervald is a Digital Worklife Strategy Consultant at Nexer with nearly 20 years of experience in Information Management. He specialises in Microsoft 365, guiding clients on how to make the most of modern collaboration tools with a focus on information security and generative AI.