Artificial intelligence: Accountability and opportunities in the insurance sphere

In a recent submission to the Treasury Select Committee, the Chartered Insurance Institute (CII) stated that institutions and individuals must be held accountable for decisions made using AI. 

It called for professionals to "always be prepared to take accountability for the outcomes created by AI, either through design or monitoring" and to be held responsible for those decisions "even if it is not feasible to explain in detail how the algorithms produce their results". The CII also stressed that this accountability must be underpinned by validation and testing to identify and mitigate discriminatory outcomes. This proportionate, regulatory approach in financial services should involve the implementation of a sector-wide strategy in which all employees of firms should receive proper training on the potential and risks of AI.

This submission has been shaped by consumer research carried out by the CII as part of its Public Trust Index which collates consumers’ and SME’s experience of buying, claiming and renewing insurance products. This research has drawn attention to the potential use of AI for support in key areas which consumers appear to value, such as 'ease of use', 'cost' and 'protection', all the while maintaining and building consumer trust. 

Dr Matthew Connell, Director of Policy and Public Affairs at the CII, said:

“While AI has been employed within insurance for many years, it is important that we continuously assess how it can be optimised for both professionals and consumers. We welcome the opportunity to offer recommendations to the Treasury Select Committee, and utilise the extensive consumer research carried out by the CII to inform this work on AI in financial services.”

Adoption of AI

There is clearly great scope for the adoption of AI in the insurance sector, particularly to speed up processes. The key challenge will be to deliver speed and service while keeping accuracy.  

In last month’s TechTalk Live webinar hosted by Insurance Times in association with Hiscox, the consensus was that AI can be used as a tool to enhance current practices. Joe Newbury, Director at Brunel PI Brokers discussed an opportunity to reform the renewal process, “We’re building a system – it’s in build at the moment – to … automate the renewal process as much as possible” he explained. However this automation will not sacrifice quality, Newbury goes on to say “This is a human-in-the-loop process where the AI is doing the heavy lifting, but the broker is ultimately checking the data to ensure quality at the end.” 

AI presents a fantastic opportunity to streamline day to day administration to free up underwriters’ time to build relationships, create value and adaptability.  “AI as a tool to enhance existing practices. It’s evolution, not revolution. We’re not replacing – we are enhancing” said Adam Atkins, head of technology underwriting at Hiscox. 

In an industry where building and keeping trust is paramount, the use of AI in the insurance sector will undoubtedly attract further regulation and guidance on its use in new products and processes. Underwriters looking to adopt AI into their common practice should consider rolling out staff training on the potential and the possible risks of AI.