AI in the energy transition: problem or solution?
Image credit: 123rf.com
The answer is both according to a panel of energy experts at London Climate Action Week
Is artificial intelligence a problem or a solution for the energy transition?
That was the question posed during a panel debate at London Climate Action Week: and the answer is… it’s both.
Or at least it is for the moment. Why? Because of the mind-boggling amount of energy it takes to power the equally mind-boggling number of data centres that are being built to enable AI.
A recent report by Goldman Sachs found that a ChapGPT query uses nearly ten times the energy needed for a Google search. Last week, Bill Gates, speaking at a conference hosted by his venture fund Breakthrough Energy, said people should not “go overboard” about AI’s massive energy demand, stating that technology companies were “seriously willing” to foot the bill for renewable energy to power their date centres.
“AI is adding to the climate problem – it has created a new challenge,” said Angie Ma, moderator of the panel discussion and co-founder of Faculty AI, a decade-old, London-based applied AI technology company that works with organisations to implement AI systems.
She said the energy transition “requires not just a transition in technology but also a transformation of the entire industry and the entire energy network”, and AI can be an enabler of that change… but first it must address the problems of its own making.
Read now: Tech Talk | AI challenges in Europe
Iberdrola launches AI innovation project to optimise power grids
Mark Butcher is founder of Posetiv, a company aiming to drive digital transformation in a sustainable way. He said that currently “we are in the ‘hype cycle’ of AI: the money is endless and the times are good”.
But “there’s a negative impact” he added, stressing how the electricity used in data centres has increased hundred-fold in recent years to accommodate computational demand.
He illustrated his point by using the case of an area in Mexico where the lights went out because all the local energy had been used up by a nearby data centre.
“AI has to help society rather than take from society,” said Butcher. And in some albeit rare case, it already is. He highlighted a UK company called Deep Green, which recaptures waste heat from data centres and redirects it to heat local swimming pools.
Collaboration key to advancing AI in energy sector
From wave energy to AI ethics: the women gamechangers in engineering
In January, Deep Green secured investment of £200m from Octopus Energy, which has built its business on the use of artificial intelligence and, more recently, generative AI.
Last month at an energy event in Lisbon, Octopus founder Greg Jackson said that generative AI has “turbocharged the power of the human”, and added that for his staff, using AI has been “like a superpower for them. It’s incredible how gen AI is not only improving external outcomes, but also internal behaviours too.” (Read more here)
AI as microcosm of energy transition
At the London Climate Action Week debate, Guilherme Castro, senior manager for energy transition and environment at Faculty, said that “what data centres are going through” was a microcosm of the wider energy transition: unprecedented change, a challenge to curb emissions and an urgency to deliver solutions fast.
“We need to think more systemically” he said, stressing the well-worn yet ever-applicable adage that the best energy is the energy you don’t use, and adding that the next-best energy you use is one that is localized.
The call for systemisation was echoed by Yujia Du, Head of Markets at Piclo, a software firm working with utilities to deliver increased flexibility in electricity grids. She said the real opportunity for AI was to help data centres operators and utilities forecast the optimum times to deliver demand response.
Beccie Drake, offshore wind digital lead at engineering and consultancy firm Arup, said she had already seen the predictive potential of AI in how it can aid wind energy planning, implementation and operations & maintenance.
However, she stressed that the energy sector was facing a chasm-sized skills gap around AI: it is going to need a huge number of data scientists if it is ever to realise the potential that artificial intelligence can bring.
AI in the energy transition will be discussed in detail at Enlit Europe in Milan in October. Register here.
Originally published on powerengineeringint.com
