AI revolution could slash energy bills as Britain urged to embrace ‘game-changing’ smart grid

Ministers were urged to act now by the report, which was described as "pushing at the absolute frontier of global thinking on deploying AI in the energy sector".

The Grid We Need Now was written by Lucy Yu, chief executive of the Centre for Net Zero and the government-appointed AI Champion for Clean Energy.

She argued that AI could help run the grid more efficiently by using real-time forecasts of electricity supply and demand, rather than relying largely on fixed, worst-case operating rules.

The UK's transition to clean energy has seen a dramatic shift in how the grid functions.

Previously, it relied on stable, continuous power from large, fossil-fuel power stations.

But increased use of renewables has led to more weather-dependent generators, dispersed all over the country.

This means balancing the grid - ensuring that supply meets demand, and that the system runs safely - has become increasingly complex.

On the old system, predicting and reacting to these needs was comparatively straightforward.

Artificial intelligence could transform the energy system within a decade, a government-commissioned report predicts

GETTY

Grid controllers could follow set safety rules and pay power stations to step in when needed.

But with so many intermittent renewables coming online, forecasting accurately has become more difficult, Ms Yu said.

"This incentivises operators to layer on additional contingency to cover uncertainty they can’t directly measure," she said.

"The result is that we are now routinely overpaying to guarantee grid headroom and services to insure against low probability events whose likelihood and consequences do not always justify the cost of full insurance."

She called for an AI-enabled system that constantly weighs up supply, demand and the risk of power cuts.

It then uses the cheapest safe combination of generation and batteries to balance the grid.

The UK's transition to clean energy has seen a dramatic shift in how the grid functions

GETTY

Experts have long sought to create such a "probabilistic" system but it is only now becoming possible, she said.

"For the first time, both traditional and newer forms of AI combined with advances in computer and access to data, enable substantial improvements to our ability to forecast and model the current and future system states, and to do so continuously at the speeds required to be operationally useful," her report found.

"Collectively, these things put true probabilistic, system-wide management of the grid within reach - a goal experts have been working towards for years."

She called for the government to "set a strategic intent to move to full probabilistic, risk-based grid operations by 2035".

"This would enable us to replace rigid and expensive-to-meet rules of thumb inputs with probability-weighted risk assessment that has the potential to deliver equivalent or better reliability outcomes at a proportionate, much lower cost", she said.

"This is a fundamental new lever now available to us to bring down high system costs. The earlier this mechanism is implemented, the sooner it can translate into critically needed and perpetual real-world savings."

Minister for Local Energy and Jobs, Martin McCluskey, referenced the report in a speech at Google's London headquarters.

He said: "The review has made a set of recommendations to government on moving towards a probabilistic grid, delivering tangible benefits for consumers, establishing clear governance and accelerating the deployment of AI solutions.

"This is exciting stuff, pushing at the absolute frontier of global thinking on deploying AI in the energy sector."

He told the Google for Startups Accelerator: Europe event: "The grid today is managing millions of decentralised energy generating assets from small to large-scale, and the only way we can handle that generation and demand in real-time is through using AI."

Alongside Ms Yu's report, the Government called for input from those in the energy sector as it looks at the opportunities AI creates while "recognising that its deployment will also present risks and wider system implications that need to be managed".

"The understanding of AI’s role in the energy system is developing, and further evidence is needed", the Department for Energy Security and Net Zero noted.