Artificial Intelligence: Emissions

Lord Bishop of Oxford Excerpts

Wednesday 23rd April 2025 Lords Chamber

Asked by

[The Lord Bishop of Oxford]

To ask His Majesty’s Government what plans they have to measure the emissions associated with artificial intelligence in relation to the United Kingdom’s net zero target.

[The Minister of State, Department for Science, Innovation and Technology (Lord Vallance of Balham) (Lab)]

I thank the right reverend Prelate the Bishop of Oxford for the Question. The Climate Change Act, as noble Lords know, made the UK the first country to introduce legally binding long-term emissions reductions targets. This sets in law our commitment to reach net-zero emissions by 2050 and means that the Government must address emissions across the whole economy to reach our targets. This includes consideration of increased electricity demand from new and growing sectors, including AI, to ensure that we are compliant with our carbon budgets.

[The Lord Bishop of Oxford]

My Lords, I thank the Minister for his Answer. He will be aware—I was not, until I researched this—that, in Ireland, data centres now account for over 20% of electricity use and the largest data centres under construction consume as much energy potentially as 2 million households, according to some estimates. Does the Minister agree that there is a need for more joined-up thinking between the Government’s energy policy and the rising energy demands of AI and data centres? As a first step, are the Government ready to accept the recommendations of the Royal Academy of Engineering for mandatory reporting for data centres on energy consumption, water consumption, carbon emissions and e-waste recycling?

[Lord Vallance of Balham (Lab)]

Recent estimates from 2022 show that data centres, including those for AI, account for about 4% of UK electricity consumption. The recently formed AI Energy Council, which is co-chaired by the Secretaries of State for DSIT and DESNZ, is set up to do exactly what the right reverend Prelate is suggesting, which is to ensure that we have a joined-up approach between energy and AI.