Leader

AI, AI….. oh

There are very occasional times when everyday life can feel like it’s something straight out of the movies.

The September 11 attacks and Covid pandemics are just some of the examples of this from recent history, where the storyline you’re living feels straight off the silver screen.

But it’s hard to think of a more a Hollywood scenario than the rise and rise of artificial intelligence.

The comments last week from a researcher at Anthropic that there is a greater than 10% chance AI could “kill all humans” within the next decade illustrate how dramatically high the stakes have risen.

In conversations with utilities professionals the topic of AI has rapidly progressed from a sense of wonder to a kind of blind panic at the risk of being left behind.

This pressure is being felt by regulators as much as by companies themselves. Not only are they struggling to keep their AI guidance relevant in such a rapidly evolving space but they are also feeling their own way through its use in their core functions.

Sources tell us that missives from regulators now often have telltale signs of having been produced with AI and there is a live debate about what role the technology will play in future price controls. Regulators have long lamented the challenge of information asymmetry when it comes to analysing the business plans of multiple companies. In theory AI sweeps this problem aside. Except of course that the companies will inevitably be deploying the technology from their own side.

It feels like some ground rules need to be set to ensure there’s a fair fight.

The use of AI in business planning might not seem very Hollywood but there are far more dramatic ways in which the utilities sector could be influenced by this digital revolution.

Last week also saw an influential report from the government’s AI champion, Lucy Yu, who recommended implementing “full probabilistic, risk-based” grid operations and planning by 2035 and 2036 respectively.

Yu also highlighted the potential benefits of agentic AI, which could be to set to work on behalf of consumers to maximise the value of their assets and lower their energy bills. However, she accepted that AI agents focused on achieving the objectives of individual consumers “may fail to deliver important collective outcomes like grid stability and resilience”.

Herein lies the complex trade off. Yes, AI can help to manage increasingly complex and unpredictable systems and find much needed efficiencies and cost savings. But utilities have a reputation for being risk averse for very good reasons. Handing control of lifeline services to a technology we do not – and may not ever – truly understand is a very, very big bet.

Of course no one is suggesting doing this without the proper checks and balances but I am concerned that leaders across the sector are facing considerable pressure to embrace the potential of AI without fully understanding the consequences.

Progress cannot be held back and history is littered with dire warnings such as this which have not aged well but I am yet to be convinced that a headlong pelt into automation is in the best interests of energy and water systems, or their customers.

But maybe I’ve just been watching too many movies.