AI is a technology that has significant potential but it also has significant environmental impact and government and businesses have a responsibility to ensure that these impacts are understood, planned for and mitigated.

The foundational models which are used in the majority of AI tools and services, require huge data processing. As larger models are required and more powerful chips are developed, higher performance is expected. OpenAI estimates that the computing resources required for AI doubles every 3 to 4 months, resulting in a 50% annual increase in energy consumption. Accommodating this additional energy requires investment and planning. 

The UK has been successful in attracting investment to build data centres across the country, but this must also be backed up by secondary investment in the infrastructure that can support the huge and growing energy requirements. Government and businesses must also recognise the impact that this will have on the local population and their access to utilities. As energy security becomes increasingly important, ensuring a continuous energy supply to data centres, upon which thousands of businesses rely, will become an issue of national importance. Furthermore, the price of energy will become an increasingly important factor for investors deciding where to invest in new data centres. Many investors have already cited the availability of huge amounts of cheap, renewable energy in Nordic countries as a reason for investing there. Building new energy sources, expanding existing infrastructure and improving efficiency and storage, are critical for the long term stability and global competitiveness of the UK AI sector.

The AI industry has a responsibility to support and prioritise energy-efficient practices and the use of renewable energy sources. This includes encouraging data centres to transition to green energy, such as wind or solar power and implementing energy-saving technologies such as liquid cooling systems to reduce electricity consumption. National and local government should consider the broader resource demands of AI infrastructure, such as water usage and explore policies that require data centres to adopt water-saving technologies and assess water usage impact, particularly in areas with limited resources. 

Addressing the environmental impact of AI hardware production is also crucial. The UK could take the lead in promoting the sustainable sourcing of raw materials and support the development of hardware that incorporates recycled or modular components, reducing reliance on rare earth metals. To tackle e-waste, both the Government and AI businesses can invest in more efficient recycling processes and promote circular economy initiatives that extend the lifecycle of hardware components.

In addition to the economic imperative, UK businesses recognise their responsibility to the environment, high-risk areas of the globe and future generations. Customers are choosing brands that recognise their environmental responsibility, often obtaining ‘B’ Corp status. Further ESG regulations such as the CSRD from the EU, place mandatory responsibilities on businesses to measure and offset not just their own carbon emissions, but the carbon created throughout their supply chain.

UK AI businesses are also part of the solution, providing innovative applications in energy management, smart city technology and agricultural optimisation that can reduce CO2 through improved efficiency. The industry should support research into more energy-efficient AI algorithms, helping businesses reduce the computational demands of AI tasks like machine learning and deep learning, ultimately lowering carbon consumption.

UKAI believes that with a coordinated approach, the UK has the natural resources and expertise to pioneer and lead the world in environmentally sustainable AI.