The Royal Academy of Engineering has warned that the rapid expansion of artificial intelligence could undermine the UK’s net-zero ambitions unless policymakers act swiftly to improve transparency, data standards, and environmental oversight of the technology.
Speaking at the First Friday editors briefing in London last week, Dr Natasha McCarthy, who leads policy at the Academy, said that while AI holds enormous potential to support sustainability and productivity, it also has a substantial and often overlooked physical footprint.
“It’s easy to think of AI as something ephemeral or purely digital,” she said. “But it depends on vast, energy-hungry infrastructure — from chips and data centres to the water and power that keep them running.”
The Academy’s recent report, Engineering Environmentally Sustainable AI, calls for stronger environmental reporting requirements and international standards to track the sector’s resource use. McCarthy noted that existing frameworks, such as the UK’s streamlined energy and carbon reporting system and the EU’s Energy Efficiency Directive, provide a starting point but remain inconsistent and poorly enforced.
“Only one in six data centres in Ireland are reporting under the current EU rules,” she said, adding that the fragmented regulatory landscape makes cross-border comparisons almost impossible.
The Academy argues that the UK could play a leading role by mandating detailed sustainability disclosures from AI companies and data-centre operators. “If we strengthen our reporting mandates and push for standardisation, we can begin to measure how much energy and water AI systems actually consume — and then improve on it,” McCarthy said.
Water use was a particular concern. McCarthy warned that some British regions are already water-stressed, yet data centres continue to use potable water for cooling. The Academy wants tougher sustainability requirements for new facilities, including closed-loop cooling systems and siting in areas with access to renewable energy and sufficient water supply. She also urged the government to link data-centre expansion with local economic development, creating “innovation hubs” that support regional businesses and skills.
Hardware design, McCarthy added, will be central to cutting AI’s environmental footprint. She highlighted research supported by the Academy into neuromorphic computing — chips that mimic the human brain and could, in theory, consume up to 1,000 times less energy than current processors. “These are not decades away,” she said. “They’re years away, and they represent a real opportunity for the UK to lead in low-energy AI.”
The Academy is also promoting “frugal AI” — smaller, more efficient models that can run on edge devices rather than relying on large data centres. Such systems, McCarthy said, not only reduce emissions but also help address data-privacy and bias concerns by avoiding massive, un-curated datasets.
Ultimately, McCarthy called for a “systems-thinking” approach that situates AI within the UK’s broader energy transition. “We can’t afford for AI to make a disproportionate demand on our electricity grid as we electrify transport and heating,” she said. “AI must have a net environmental benefit — enabling decarbonisation, health innovation, and scientific discovery — not undermining them.”
There’s plenty of other editorial on our sister site, Electronic Specifier! Or you can always join in the conversation by visiting our LinkedIn page.
