Gen AI terrain ‘marked’ with ‘excitement’ and ‘apprehension’

Millions of pounds are being invested into implementing generative artificial intelligence (GAI) technology in the UK insurance industry, but opinions remain “deeply divided” on digital development, said FintechOS cofounder and chief executive Teodor Blidarus.

According to the firm’s new study – entitled Generative Artificial Intelligence: The Technology Polarising the Financial Services Industry, which was published on 22 November 2023 – 44% of UK insurance institutions are already using GAI today, with 48% investing in GAI research and another 39% in its implementation.

Costwise, 35% of firms have invested between £800,001 and £1.6m, with a further 27% investing between £1.7m and £4m.

And because of GAI investment, 66% of UK insurers anticipate an average 20% revenue increase within the next three years, while 29% and 37% predict a rise in revenue of between 11% and 20% and 21% and 30% respectively.

Blidarus highlighted: “GAI is a terrain marked by both excitement and apprehension.”

“While opinions within the insurance industry are deeply divided, there is one common consensus – the expectation that GAI will boost revenue but inevitably reshape the workforce and displace jobs”.

Research company Censuswide surveyed 1,000 senior technology decision makers at insurers and major banks evenly distributed across the UK and United States during August 2023 for the report.

The findings further revealed that insurance executives believed employee upskilling will be the business area that will benefit most from GAI (24%), followed by pricing optimisation (23%), customer credit checks (22%) and claims assessment automation (22%).

Some 21% noted that they believed the impact of GAI would accelerate development timelines and 17% said it would reduce workloads.

However, at the same time, around 25% of respondents described themselves as frightened of GAI, while 23% admitted to feeling curious.

Over half (60%) of respondents also stated that GAI would result in job losses – 36% predicted that this would equate to between 21% and 30% of the workforce over the next three years.

‘Low-hanging fruit’

Speaking exclusively to Insurance Times, CFC’s chief underwriting officer Andy Holmes said that the insurance industry’s “use of any technology, whether it’s artificial intelligence or otherwise, is not going fast enough”.

“[Lloyd’s modernisation programme] Blueprint II, if it was done tomorrow, would be fabulous, because there’s still too many processes and too much complexity for the type of insurance that we do,” he added.

While AI is a broad field of computer science focused on creating systems that would typically require human intelligence, GAI more specifically describes a particular field of AI that encompasses image generation, virtual chatbots or machine learning models 

A spokesperson at Insurance Times’ Broker CEO Forum 2023 retreat, held on 16 and 17 November at Lainston House, Winchester, highlighted that firms first need to “focus on the foundations” when it comes to GAI.

“I wouldn’t jump straight to [GAI] and think it’s going to transform your business, because it won’t. You’ll need traditional machine learning [and] AI,” they explained.

“There is so much low hanging fruit in the machine learning space. I would almost wait until generative AI matures to a reasonable degree before going all in or run it in parallel – more as an innovation in [research and development] piece.”

‘Could we, should we?’

FintechOS’ study additionally found that when respondents were asked whether they viewed GAI as friend or foe, 42% opted for friend, with 46% stating foe.

UK insurance executives said the most prominent concern causing the most apprehension was GAI’s potential for discrimination and bias (26%), followed by changing public opinion of the technology (25%) and a lack of data transparency (25%).

At Insurance Times’ roundtable, held in association with Shift Technology on 14 November 2023, LV=’s director of financial crime Ben Fletcher highlighted that there seems to be a “disparity” between “people’s expectations as a consumer versus our views of what people want” in terms of AI use.

He said that consumers’ perspectives of the large amount of data the industry sits on to inform decisions contrasts with the “risk adverse view” of the insurance sector.

“At the moment, there’s a big focus on how do we make use of the tech – the ethics conversation of ’could we, should we’ is going to escalate quite significantly.”

Katie Davies, head of underwriting services at Ageas, also highlighted that firms need to approach using AI during decision-making at the claims stage in terms of fraud with caution.

“We really have to be careful that we’re using this as a prompt [but still have the] human judgement to say, actually this is more likely to be a fraudulent case, therefore let’s apply some care to it rather than say ’sorry, computer says no, we’re not paying your claim’,” she said.