Energy use concerns dampen businesses’ interest in AI for sustainability

Nearly three quarter of businesses worry that AI's massive energy consumption may outweigh its benefits in supporting sustainable goals.

In a survey of 1,300 business leaders across industries in Asia, Europe and the Middle East, commissioned by Alibaba Cloud, 76% said they are “intrigued” by the potential of artificial intelligence in driving sustainability – but 61% admitted they were concerned about AI’s energy use.

In fact, as many as 71% of businesses believe that the substantial energy consumption of data centres needed to power AI may outweigh its benefits for sustainability.

This suggests that AI’s environmental impacts may be a barrier to adoption, even as tech companies promote its potential to accelerate the climate transition. 

Sustainable AI: A growing movement

Digital sustainability experts have been warning about the massive energy and water consumption needed to fuel AI models for years – with data centres expected to take up 12% of the entire US electricity generation by 2028. 

The conversation appears to be now reaching the mainstream, with governments forming sustainable AI coalitions and tech firms launching AI energy efficiency benchmarking platforms. 

The movement for sustainable AI has gained strength since Big Tech firms disclosed their first emissions data since the explosion of AI: Microsoft reported a 30% emissions rise in 2023 alone, while Google’s emissions grew by 13% the same year, and 48% since 2019.

AI seen as a priority for sustainable development

Despite their concerns, many companies in the survey identified AI and digital technologies as crucial for their sustainable transformation. AI and machine learning are particularly viewed as crucial digital technologies for advancing corporate sustainability in Middle East firms (52%), while Europe (41%), emerging Asian markets (40%) and developed Asian markets (36%) place less emphasis on it.

However, most businesses worldwide (81%) do believe human oversight is needed to guide the development of AI tools.