Image source, Getty Images
- author, Imran Rahman Jones
- role, Technology reporter
-
Google’s latest environmental report found that the company’s greenhouse gas emissions in 2023 will be 48% higher than in 2019.
The tech giant blames the rise in the amount of energy data centres need, exacerbated by the explosive growth of artificial intelligence (AI).
AI-based services require significantly more computing power (i.e. electricity) than standard online activities, sparking a series of warnings about the technology’s environmental impact.
Google’s goal is to achieve net-zero emissions by 2030, but it acknowledges that “further integrating AI into our products may make it harder to reduce emissions.”
Data centers are essentially collections of vast numbers of computer servers, and AI requires vast numbers of them.
A recent study found that generative AI systems such as ChatGPT can consume around 33 times more energy than machines running task-specific software.
But Google’s report also revealed significant disparities in the impact of its data centers around the world.
Most centres in Europe and the Americas derive the majority of their energy from carbon-free sources.
In comparison, data centers in the Middle East, Asia and Australia use significantly less carbon-free energy.
Overall, Google says about two-thirds of its energy comes from carbon-free sources.
“When you actually go into a data centre, it’s very hot and very noisy,” says Tom Jackson, professor of information and knowledge management at Loughborough University.
“People don’t realise that everything they store in the cloud impacts their digital carbon footprint,” he says.
Professor Jackson runs the Digital Decarbonisation Design Group, which aims to measure and find solutions to reduce carbon emissions through data use.
“Data providers need to work more closely with large organisations to help them stop storing huge amounts of dark data,” he says.
Dark data is data that has been collected by an organization but has never been used or is never used at all.
But storing it on a chip consumes a lot of energy even when it’s not in use.
“On average, 65% of the data stored by organizations is dark data,” says Professor Jackson.
He praised Google’s goal to make its data centers net-zero by 2030, but said it would be “really tough.”
AI’s increasing energy and water usage has sparked a wave of warnings, especially as the sector is predicted to continue growing rapidly in the coming years.
The head of Britain’s National Grid said in March that the combination of AI and quantum computing will drive a six-fold increase in demand over the next decade.
However, Microsoft co-founder Bill Gates recently downplayed the environmental impact of AI.
Speaking in London last week, he suggested that AI could increase electricity demand by between 2 and 6 percent.
“The question is, will AI accelerate reductions beyond 6 percent? The answer is, absolutely,” he said. Financial Times.