One of three research projects commissioned by Business Secretary Greg Clark will look into how firms can leverage artificial intelligence in insurance, and an executive has offered insights as to how to undertake the initiative.
“It’s great news that the government has turned its gaze to technology in the UK insurance sector as this should help spur on a traditionally conservative industry,” commented Tony Tarquini, European insurance director at Pegasystems. “By their very nature, insurance companies have always been rather risk-averse and slow to adopt the latest technology, and this extends to the uptake of AI.
“Given this mindset, during these projects the government must be wary to not thrust change upon the insurance industry, but develop a constructive forum for insurers about how to best apply AI in a safe and practical way that will truly add value.”
The probe will examine how the technology can be applied to areas such as underwriting and claims processing, with the hope of transforming delivery of insurance services and bringing about savings for consumers.
“AI has been around for a long time, with reams of research, theoretical PhDs, etc., available on the topic, but in terms of the day-to-day operation of an insurance company, insurers really struggle to apply AI technology at a basic level,” added the Pegasystems official.
“Yes, there are some insurance companies who are already taking their tentative first steps into the world of AI, but many are not, and these are the organisations who are putting themselves in danger of missing out on massive returns and succumbing to the onslaught of insurtech innovations. At an operational level, the key to its successful implementation is determining the best means of applying AI in a real office environment.”
Meanwhile he also pointed to the highly regulated environment in which the insurance industry operates, suggesting the need for ethical and transparency parameters for what was termed as ‘responsible AI’.
“Other industries have seen AI amplifying inherent prejudice and this has to be avoided when writing policies and judging claims,” noted Tarquini. “Furthermore, the transparency of the AI in use has to be at a level whereby a regulator can understand the rules and algorithms in place which produce a set of specific outcomes.”
