Artificial intelligence (AI) has arrived in public consciousness with astonishing speed. In just a few years it has moved from being a specialist technology discussed quietly by software engineers into something used daily by students, businesses, governments and ordinary families.

AI can now write reports, create images, analyse data, answer questions and increasingly automate tasks which once required highly trained people.

To many, it feels almost magical, invisible, weightless, existing somewhere “in the cloud”.

But there is a problem with the way we talk about AI. We speak about it as though it is purely digital, when in reality AI is one of the most physical technological revolutions the world has ever seen.

Kevin Keable, Chair of the EEEGR (Image: Sonya Duncan)

Every AI generated answer, every search, every image and every automated process relies on vast amounts of infrastructure sitting firmly in the real world. Huge data centres filled with servers, cooling systems, backup generation, high voltage substations, transmission systems, fibre optic cables, engineering teams, construction workers, maintenance technicians, power stations.

And above all else: electricity. The growth of AI is already beginning to reshape global energy demand forecasts. Major technology companies are investing hundreds of billions into AI infrastructure and governments are now discussing sovereign computing capability in the same way they once discussed oil reserves or industrial manufacturing.

The challenge is that this revolution is arriving at precisely the same time as another enormous transition: the electrification of society itself. We are already trying to increase electricity generation to support electric vehicles, heat pumps, industrial decarbonisation and new manufacturing.

Now AI is adding another layer of demand on top. Data centres alone are becoming enormous consumers of electricity. Recent discussions in Essex highlighted that planned data centre developments could require electricity demand equivalent to well over current domestic consumption levels in some areas.

Similar pressures are emerging across Europe and the United States. The “cloud” turns out to have a very large appetite for power.

This matters because energy infrastructure cannot simply appear overnight. Building generation, upgrading transmission networks and expanding industrial capability takes years, often decades.

The UK learned this during the transition from coal gas to natural gas in the 1960s and 70s, when millions of homes, businesses and appliances had to be converted in one of the largest engineering programmes in British history.

The AI revolution may demand a similarly ambitious response. Yet perhaps the greatest challenge is not technology. It is people.

Who will build and maintain this infrastructure? Who will install offshore wind farms? Who will weld the pipelines, fabricate the substations, upgrade the grid, maintain the gas systems that still provide flexible generation, design cooling systems for data centres and operate future hydrogen and carbon capture projects?

At present, the UK and particularly this region already faces shortages in many technical and engineering disciplines. Across the energy industry there are growing concerns about shortages of welders, HV technicians, process engineers, fabrication specialists and experienced project personnel.

Colleges and training providers are struggling to recruit lecturers with industry experience while employers compete for an ageing workforce. This creates what could become a dangerous contradiction.

Young people are increasingly encouraged to see their future through the lens of software, apps and AI tools, while at the very same moment the world may require one of the greatest expansions of physical infrastructure and engineering capability in modern history.

In simple terms, if we do not inspire and train enough young people to build the infrastructure behind AI, we risk creating a double whammy for the next generation: AI replacing some jobs while too few young people gain the skills needed to build the energy systems, data centres and infrastructure that AI depends upon.

This is not an argument against AI. Far from it.

AI has the potential to transform productivity, improve healthcare, accelerate scientific discovery and optimise industrial operations, including within the energy sector itself. AI is already helping operators improve maintenance planning, analyse offshore assets, optimise energy systems and enhance safety performance.

But we must be realistic about the scale of what sits beneath it. Artificial intelligence will not run on enthusiasm alone. It will require affordable, reliable and scalable energy systems backed by long term infrastructure planning and a skilled workforce.

That is why regions like the East of England matter enormously. This region already plays a central role in the UK’s energy system. Offshore wind, gas infrastructure, interconnection, ports, engineering supply chains and nuclear capability are all concentrated here.

The East of England is not peripheral to the national energy conversation – it is one of the places where the future will physically be built. And perhaps that is the most important lesson of all.

For years Britain has often separated the digital economy from traditional industry, as though coding and engineering exist in different worlds. In reality, they are now converging rapidly.

The AI economy will depend not only on programmers and software designers, but also on electricians, fabricators, project managers, planners, process engineers and construction teams. The future will belong to countries capable of combining both.

Because behind every AI generated answer is something very real: concrete, steel, cables, cooling systems, substations, power stations and skilled people.

The nations that understand that connection earliest and invest in it properly are likely to lead the next industrial era.

This story is also published in Insight Energy magazine, covering the latest news from the UK’s energy sector. Read the latest edition here.

Source link