In light of the rapid growth of artificial intelligence (AI) technologies, a pressing call has emerged for legislation requiring tech companies to disclose their data centre energy and water consumption. Experts warn that this surge in AI activity poses a significant risk of environmental damage.
A report from the National Engineering Policy Centre (NEPC) highlights the unprecedented acceleration of AI systems and the environmental implications that come with it. Professor Tom Rodden, pro-vice-chancellor of research at the University of Nottingham and contributor to the NEPC, emphasizes the urgency of implementing mandatory reporting. He notes that AI advancements have led to an exponential demand for computational power, often at the cost of resource efficiency. This trend could result in irreversible harm to our planet if not addressed.
Significant year-over-year increases in water consumption by major tech players such as Google and Microsoft further illustrate the issue at hand. Reports indicate that Microsoft’s data centres consumed approximately 6.4 million cubic meters of water in 2022, marking a 34% rise from the previous year. Similarly, Google reported a water use increase of 20%, totaling 19.5 million cubic meters. Such figures raise concerns, particularly as many withdrawals are sourced from potable water supplies.
One of the fundamental challenges is the lack of reliable data on resources consumed by data centres. The NEPC report advocates for the government to develop a comprehensive framework for gathering this data, which is essential for formulating effective policy measures aimed at curtailing energy consumption.
Rodden underscores the importance of tracking AI’s environmental costs. He asserts that access to accurate data regarding environmental impacts will facilitate better targeting of efficiencies in AI development and deployment. This approach is crucial for ensuring a sustainable future for AI in the UK.
The NEPC report outlines several recommendations for policymakers, including:
Professor Sarvapali (Gopal) Ramchurn, a key figure in the discussion and chief executive of Responsible AI UK, stresses the need for transparency regarding the environmental consequences of AI technologies. With AI’s role in enhancing productivity and innovation over the last two years, it is crucial for providers to be forthcoming about the associated environmental costs. As Ramchurn aptly points out, without measurable data, managing these impacts and ensuring mutual benefits becomes a daunting challenge.