By Tatiana Gaitan, graduate consultant, Highbury Communications
The UK Government’s AI Opportunities Action Plan, announced this week, aims to ramp up the adoption of artificial intelligence (AI) across the economy. However, the lack of focus on sustainability raises concerns. The AI Energy Council and future recommendations outlined in the AI Action Plan begin to address some of these issues, but more significant effort is required. This includes increased investment, a deeper understanding of the “green elephant” in AI development, and a proactive approach to addressing the environmental challenges of the technology now and in the future.
This week, the government revealed its commitment to take forward 50 AI recommendations from a UK action plan for supporting the growth of AI led by Matt Clifford, a tech entrepreneur and Chair of the Advanced Research and Invention Agency.
Keir Starmer, Prime Minister of the United Kingdom, announced plans to adopt these recommendations throughout 2025. The government’s plan highlights the necessity for a strong AI infrastructure, promising to upgrade the National AI Research Resource’s capacity and create AI data centres with power outputs ranging from 100 MW to 500 MW. Additionally, it will establish an AI Energy Council to tackle emerging energy requirements in the UK.
Despite the creation of the AI Energy Council, only recommendation five addresses sustainability challenges in AI infrastructure:
“Recommendation 5: Mitigate the sustainability and security risks of AI infrastructure, while positioning the UK to take advantage of opportunities to provide solutions.”
“Government response: Agree. DSIT will set out how the UK will seek to address the sustainability and security challenges of AI infrastructure as part of its long-term compute strategy.”
While the council is a promising start, experts argue that AI’s environmental impact extends beyond energy consumption. The lifecycle of AI technologies, from the pre-modelling, modelling, and post-modelling phases, has challenges that remain largely unaddressed in the plan.
For instance, the production of graphic cards, which are crucial for generative AI operations and lead to greenhouse gas (GHG) emissions, according to research from Capgemini. AI systems consume significant amounts of energy and water during the modelling phase. At the same time, by the end of its lifecycle, generative AI is estimated to produce between 1.2 and 5 million tons of e-waste by 2030.
Some experts have called for equal investment and ongoing efforts to confront these environmental issues. The director of the Ada Lovelace Institute, Gaia Marcus, emphasised the urgency of tackling the “negative impacts of AI both now and in the future,” urging the government to allocate resources wisely. The Green Party labels the AI strategy’s environmental considerations a “green elephant in the room,” highlighting undiscussed environmental harms by AI technologies.
Likewise, the Royal Academy of Engineering approved the establishment of the AI Energy Council as a solid partnership and a chance to address the pressing clean energy crisis, which is categorised as “immediate”.
While the AI Action Plan represents a significant initiative for the UK Government and its attempt to boost growth by investing in and adopting AI, it also needs to consider the broader environmental implications of the technology. The AI Energy Council and recommendation five acknowledge these challenges, marking only the initial steps towards addressing the adverse effects that AI development has on the environment. Every stage of AI’s lifecycle contributes to climate change, making increased investment and awareness of this issue essential.