by Anton Clennar, Associate Director, RNA Analytics
Since the launch of ChatGPT at the end of 2022, artificial intelligence has captured the public’s imagination as well as investors’ attention. Over the last twelve months in particular, companies around the globe have begun to experiment with all forms of AI, some desperate to be the first to tap into the raft of rapidly maturing technologies across the AI sphere. For some industries, AI has the potential to be quite transformative – particularly those for which data plays a pivotal role.
This ability to tap into data – at speed – is likely to be a critical differentiator in the insurance industry, where incumbents are working hard to identify opportunities – albeit at a more ‘risk-managed pace’ than their counterparts in banking.
Recent years have seen the costs of implementing AI come down significantly, as the market has matured – suggesting even better value for insurers as they seek to explore the benefits across all segments of insurance provision.
According to Goldman Sachs’ Global Insurance Survey 2024, some 29% of insurers are currently using AI, with 51% looking to implement some form of AI technology soon. Insurers see AI as having a broad range of uses, with 73% either using it, or considering using it, to reduce operational costs; and 39% using or considering using it in underwriting. Other uses cited in Goldman Sachs’ research include claims management, investment evaluation, and, more visibly, ameliorated client service functions.
This is especially powerful for insurers engaged in high-volume personal lines business, where less human interaction is necessary than with the large, complex risks and claims associated with property/casualty lines.
For actuaries across all segments of insurance, AI can help process and analyse data more efficiently than ever before. Whilst it won’t replace actuaries, AI will transform their day-to-day work, making them more agile and effective. AI-powered predictive models estimate potential losses, aiding insurers in decision-making, allowing actuaries to focus on higher-value tasks. Routine tasks such as data pre-processing, model fitting and report generation can, with the right tools, be largely automated – freeing up actuaries’ time for more strategic analysis.
As with all technologies, whilst AI is exciting, the models introduce new sources of uncertainty for actuaries, and human analysis is vital to the avoidance of biased, unethical or inaccurate predictions or models – particularly in pricing and underwriting. For this reason, data and models must be regularly audited for bias, and actuaries should closely review model outputs to ensure fairness.
Fairness and ethics have been the focus of much discussion around the industrial implementation of AI globally, and it is important that time is taken to get this right. Central to this is an acknowledgement that AI should be used to augment human capabilities, and not to replace workers – especially where human creativity, values and strategic thinking are necessary – such as across the entire insurance value chain.
Adoption of AI should reflect a consideration of these basics. At the time of writing, a group of employees at a number of well-known artificial intelligence companies issued an open letter warning of a lack of safety oversight within the AI industry – the very oversight that AI companies themselves acknowledge is necessary to fair outcomes in AI use.
Looking ahead, some commentators predict that AI could have a transformative power similar to that of cloud computing and semi-conductors. Much work and debate will be needed to advance with AI at the right pace, and gain the competitive advantage it promises.
