The Lloyd’s Market Association (LMA), in partnership with Barnett Waddingham, has published a survey examining the use of artificial intelligence (AI) and machine learning (ML) in actuarial and risk functions.
The findings suggest that while interest in these technologies is growing, adoption remains limited due to a range of practical and regulatory concerns.
The report, AI and ML in Actuarial and Risk, outlines issues cited by respondents, including difficulties in validating outputs, concerns about accuracy, and a general lack of internal skills. Many organisations said additional training and clearer guidance would be necessary before these tools could be applied more broadly.
Some respondents expressed reservations about using third-party AI tools, particularly in relation to compliance requirements. Uncertainty around regulatory expectations was described as a factor slowing uptake, with many firms opting for a cautious approach.
Meanwhile, attitudes towards AI and ML differed by function. According to the report, actuarial professionals were found to be generally more open to using these tools, noting the alignment with existing data-driven processes. Risk professionals appeared more reserved, often working with less structured data and raising concerns about reliability and oversight.
Data quality was identified as a key issue, with respondents highlighting the importance of transparency, explainability, and regular model updates. While data limitations were acknowledged, the report suggests that both human and machine systems can adjust over time to these constraints.
The report also found that current use of AI and ML appears limited. Several organisations reported early-stage experimentation, such as using AI to support claims classification or detect pricing trends. About half of respondents indicated minimal or no current use of these tools in their day-to-day work.
The findings reflect broader trends in the UK insurance sector, where adoption of AI has been gradual.
Although some insurers have started applying AI in underwriting, claims, and fraud detection, implementation across the market remains uneven. Regulatory guidance from the Prudential Regulation Authority and the Financial Conduct Authority has urged firms to assess AI applications carefully, particularly in relation to transparency, fairness, and governance.
“The results of this survey show that there is a long way to go in adopting and leveraging the full potential of AI and ML tools,” said Sanjiv Sharma, head of actuarial and exposure management at the LMA. “Currently, the focus is on automation and efficiency gains, but I would encourage market firms to actively explore the opportunities in this space.”
Sharma added that firms that successfully integrate AI and ML into their operations can gain a competitive edge and make more informed strategic decisions.
“They are tools that can drive real innovation, if users are able to balance the potential benefits with the compliance and ethical considerations that will be key to Lloyd’s participants continuing to lead in global insurance,” Sharma said.
Meanwhile, Wan Hsien Heah, partner and head of general insurance at Barnett Waddingham said: “Challenges remain when it comes to further adoption but the survey will hopefully spur discussions that will lead us towards defining what market practice looks like for implementing AI and ML in actuarial work within general insurance.”
The survey drew responses from organisations representing around 55% of the market’s stamp capacity and included six in-depth interviews.
