The insurance industry is undergoing a significant transformation—and it’s all thanks to the rapid evolution of modern technology. What was not so long ago, a sector very set in its ways to do things in a traditional, manual way, is now recognising and leveraging the power of tools like AI. In fact, recent research has shown that the global AI in insurance market accounted for $8.13bn in 2024 and is projected to surpass $141.44bn by 2034.  

Not only are insurance businesses using it to accelerate repetitive paperwork tasks, but it’s also extending to many of its core processes as well—from increasing efficiency of claims processing and customer service resolutions, to enhancing compliance. 

But, with so much discussion around the implications of AI on humanity, it’s crucial to think about the ethical considerations involved when investing in the technology. For this month’s blog, I sat down with Paul O’Brien, who is our Group Chief AI Officer, to discuss how sustainability, AI and ethical practices are all inextricably linked.  

Hi Paul, thanks for joining me! You were appointed as Davies’ Group Chief AI Officer last year, to help us advance our use of AI, ethically. Can you talk us through your strategy and how this aligns with sustainability?  

Paul: Hi Gillie! No problem at all, thanks for having me. Well, much like your People, Planet, Purpose strategy, my ethical AI strategy has three core elements at the root of it. First, we need to make sure that we are using AI in a way that aligns with ethical and regulatory standards. To do so, we’ve established AI usage guidelines which outline our AI framework and our acceptable AI usage policy. Our framework is aligned against the EU AI act which focuses on using AI safely within business, and meeting legal obligations around fairness, transparency, and navigating ethical concerns as well.  

Our second pillar is anchored around how we will promote responsible AI use throughout the organisation. So, we are focused in areas like data security and customer privacy throughout our business. We have a process in place called the AI impact assessment which our AI and governance teams use to risk assess new products and services within Davies that use AI, before they are released into the live environment. This means, we’re not just creating new technology and letting it loose into the world—we identify and work to mitigate any risks that could be a gateway to a much bigger problem or issue, this might be raising our carbon footprint for example.  

The last element of the AI strategy looks at how we can develop a workforce within Davies that is skilled in AI. So, I have a small AI team—there’s just five of us now effectively working within it—and our approach is not just to use AI ourselves, but to look at how we can embed AI throughout the business and how we can give everyone an opportunity to leverage it. To help with this, we’re looking at what training we can deliver to the whole business to turn it into a workforce skilled in AI.  

In this blog series, we’ve previously spoken about how we’re currently focused on building increased sustainability in our service delivery, and how we’re looking to doing more of that this year. But when it comes to AI, there’s certainly been wider industry concerns over its ability to be sustainable. What do you say to those who hold these concerns? 

Paul: I think it’s a very justified concern when you approach it at face-value. To train an AI model and get it ready for the things you want it to do, you need to use a lot of power and do mathematical computations that require a large number of expensive GPUs (graphics cards).  

With that said, we have a responsibility to leverage evolving technology for the benefit of Davies, while understanding and balancing these concerns. AI is becoming much more efficient and we’re finding much better ways to train models and advance AI without it being quite so costly, both spend-wise, and to the environment. For example, there’s recently been a Chinese AI app in the news called DeepSeek which is a low-cost AI model providing conversational AI that costs just a fraction than some of the major players of the tech world.  We predict more and more businesses will begin to think about massively reducing the energy usage of their own models, making the industry more energy-efficient in relation to AI usage.  

Additionally, by giving our models the ability to innovate and streamline laborious processes, where previously it might’ve meant a person having to do it manually, or more site visits required, we are able to reduce environmental impact this way. At Davies, we’ve been able to leverage the power of technology to build sustainability in our property claims service delivery, and are committed to doing it on a much wider scale this year.  

So, my final take is that, while it is a justified concern, I have a strong degree of confidence that we can leverage developing technology to be more energy efficient in the long run.  

You touched on this slightly in your first response, but one of the main concerns raised when talking about AI from a sustainability viewpoint is that it contributes to higher carbon emissions. What do you think needs to be done in the future to help mitigate this?  

Paul: A couple of things, really. When you’re training all these models and running AI workloads, they’re in large data centres owned by the likes of Microsoft and Amazon, and we need to ensure they are continuing to drive sustainable energy use within the data centre. I think it’s important to understand that the energy is being used to generate our AI solutions and AI results is renewable, but also to remember that just because it’s renewable, it doesn’t mean we want to use as much of it as possible. We still need to be mindful and responsible about what we’re doing. Similarly, the continued optimisation of technology can help. We’re constantly striving to make models and AI generally better, but the tech industry also needs to consider making it more efficient as well.   

Although AI seems quite developed already, do you see any limitations in current models that need to be addressed to unlock their full potential in driving sustainable outcomes? 

Paul: Current large language models—like ChatGPT—cost money and use a fair bit of energy to train. So, conversely, using small language models have less of an impact both financially and environmentally. By being more mindful about picking the right model for what is being done in AI will help us to become more sustainable and efficient and to drive carbon emissions down.  

Thinking about a real-world example, Microsoft are doing a great job of this currently. They have several tools within their platforms where, if you’re using Microsoft Azure, you can view what your energy usage is for that AI, before you deploy it or when you’re using it. I think across the AI landscape it would be great to have that sort of visibility—to be able to see the real-world impacts.  

That’d certainly be something to look out for! Speaking of the future, let me wrap up by asking you this: What emerging AI technologies or trends do you think will have the most significant impact on global sustainability efforts? 

Paul: For me, it’s that just the super-fast pace of innovation in AI is no longer focused on getting better results—it’s focused on becoming more efficient, both in terms of costs but also consequentially, sustainability. As we’ve mentioned, AI can be quite expensive to use, and people are trying to drive costs down, and ultimately, this means making it use and require less power.  

So, this driving down of cost is making it more efficient and that’s where the market is heading. AI has a twin track of being more efficient, and therefore cheaper and more advanced at the same time—it’s not a case of one or the other. I think this is really encouraging for people who worry about the impact of AI on the environment, and I predict we’re going to be seeing much more of it.  

With the future of AI looking up, both in terms of function and sustainability, now is the time for all businesses to be considering their AI usage and the ethicality of it. To hear more of Paul’s insights into ethical AI, you can catch him over on his LinkedIn or visit us on the Davies’ website to find out more about our Responsible Business strategy.