The rose-tinted glasses through which many big insurers and reinsurers have viewed the potential of introducing generative AI into their business models are being laid aside and a new perspective is taking hold.
AI, large language models and big data offer great hope. But they also offer big energy bills and that tends to be bad news for planet Earth. The (re)insurance community is very much at the forefront of discussions about climate change because, more often than not, they have to pay for the damage.
How does today’s large (re)insurer square their love affair with AI and its potential for their business with their climate commitments and net-zero targets?
One option is to look more at alternative energy. We have already heard of the efforts of the biggest tech companies – Amazon, Google, Meta – to act on this because they crave huge amounts of low-carbon 24-hour electricity to power their data centres. Otherwise, how can they win the AI war?
Earlier this year the Electric Power Research Institute, a not-for profit body that works to advance the clean-energy transition, forecast that data centres could double their share of US electricity use by 2030.
The suggestion is that big tech will build their data centres close to power-generating facilities – probably nuclear-fuelled – that can provide such energy. The reality is that there are not yet enough nuclear-powered plants to go around – and for good reason.
Data centres typically make their money because of speed to market – and what is clear is that new nuclear facilities are not speed to market. They are typically plagued by project timeline overruns as well as cost overruns. And they are enormously risky for whoever is funding them.
Already national climate-change targets are coming under close scrutiny because of the increase in AI use. As an indicator, S&P Global Commodity Insights has reviewed its forecast for coal plant closures downwards by 40% by 2030.
China looks like it might be a leader in the field of green-energy transition, committing $800bn until the end of the decade to shift energy production from coal to renewables.
Where all of this leaves the (re)insurance community is not clear. Companies cannot afford to take a ‘wait-and-see’ approach to using AI unless their competitors are doing likewise.
Simply refusing to take the lead and wait for guidance and examples from the big tech companies is also something of a cop out – especially for an industry that has been navel-gazing about climate change for the best part of a decade.
It may very well become a defining issue when looking to recruit the next generation of workers for the sector.
No one has a neat answer to this problem yet, but it certainly should rise to the top of the risks that the industry is discussing – both in public forums and in the boardroom.
It is also a dialogue that must involve national governments and, very probably, the insurance regulatory community. A
Paul McNamara
Editorial director
Asia Insurance Review