AI in energy: how assurance can support responsible deployment

AI could deliver substantial benefits for energy consumers and the wider energy system, including better customer service and forecasting, more efficient network operation and asset management, and better-informed trading decisions.

Some AI applications also have a higher risk profile. Their behaviour can be uncertain, their decisions can affect many people or transactions quickly, and some applications can take consequential actions without prior human approval.

In our response to Ofgem’s Call for Input on AI Assurance, we argue that deployment optimisation can provide a practical link between governance principles and the decisions firms must make about how individual AI applications operate.

Deployment optimisation means comparing feasible ways of configuring and controlling an AI application, then selecting the approach expected to produce the greatest overall value for consumers and the energy system. This requires firms to consider the benefits, downside risks and costs of each option within existing legal and regulatory constraints.

The approach makes choices that might otherwise remain implicit more explicit. Firms need to decide what an application should achieve, what information it may use, what actions it may take, when people should intervene and how performance should be monitored.

Responsible deployment does not mean seeking to eliminate every conceivable risk. Additional controls may reduce potential harm, but they may also limit access, speed, accuracy or other benefits. The task is to identify the best feasible approach while meeting legal obligations, managing risks appropriately and considering how benefits and harms are distributed.

These decisions are also shaped by regulatory expectations. An AI application may create substantial benefits overall while still leaving some consumers worse off. If firms are uncertain how Ofgem will assess that trade-off, worthwhile investments may not proceed, particularly where firms cannot capture all the wider benefits created for consumers and the energy system.

Ofgem can therefore do more than set constraints. It can help create the conditions for responsible adoption by clarifying how deployment decisions will be assessed, setting proportionate expectations for evidence and identifying risks that may only become visible across the sector. Greater clarity can strengthen consumer protection while giving firms more confidence to invest in applications that could create value.

Read Frontier’s full response to Ofgem’s Call for Input on AI Assurance.