How green is your AI?

Artificial Intelligence has a sustainability problem - here's why

- DAVID CROOKES

For years, Google has been an important tool for anyone looking to find answers on the web, In recent times, however, artificial intelligence (AI) has also begun to play a key role. Search Google today and you’ll likely see an AI overview delivered using the tech giant’s custom Gemini model which is capable of multi-step reasoning to help with even the most complex of questions. More and more people are also turning to Microsoft Copilot and ChatGPT but such actions are not entirely without consequence.

According to the International Energy Agency (IEA), a ChatGPT query requires nearly 10 times as much electricity to process as a standard Google search: some 2.9 watt-hours (Wh) of electricity compared with 0.3Wh. It’s a startling statistic which leads to questions about the sustainability of artificial intelligence given the sheer number of people and businesses that are making use of AI.

Figures published by Statista show that the number of people using AI tools surpassed 250 million last year – double that in 2020. Forecasts also suggest growth will see the figure hit 700 million in six years time.

It’s easy to see how AI will continue to enjoy an upward trajectory given the sheer enthusiasm among Big Tech companies. Apple Intelligence, for instance, is set to become an integral part of iPhones, iPads and Macs for years to come, allowing you to create images, better engage with Siri and enjoy new writing tools for more effective communication. It’s already available in the US and it’s coming to the UK, Australia, Canada, New Zealand and South Africa by the end of the year. But it stands to reason that the more AI is adopted, the greater the potential strain on resources.

A powerful problem