The question that many insurance industry CEOs will be mulling over at present is whether generative AI is going to be a game changer for the sector – or just another flash in the pan.
While some newspaper headlines have gone for the shock value of suggesting that generative AI will mean the end to 30% of jobs currently undertaken by humans – presumably including jobs in the insurance sector – the reality is likely to be more nuanced.
Zurich was the first global insurer to indicate that it was going to experiment with ChatGPT, a chatbot developed by OpenAI and launched in November 2022. The insurer indicated that its first experimental applications of the technology would be in claims and writing code for statistical models.
This is not Zurich’s first attempt to harness AI technologies in its core business. At the end of 2021 the insurer agreed to acquire Estonia-based company AlphaChat, which provides conversational AI technology for customer service automation to enhance the group’s digital capabilities.
At present, institutions toying with the new technology are suggesting that it will help workers become more efficient rather than replace them entirely. But the reality is that no one yet knows where it will lead.
Shortly after Zurich’s ChatGPT announcement, another Swiss-headquartered insurer, Helvetia, said that it was the world’s first listed insurer to launch a direct customer contact service based on ChatGPT.
The service uses AI to answer customers’ questions on insurance and pensions. “Although currently still at the experimental stage, the long-term hope for the service is that it makes access to insurance and pension products more straightforward,” a statement from the insurer said.
Helvetia has gone to some lengths to emphasise that it wants to gauge the potential offered by language models and gain insight into how customers perceive such services.
Helvetia Group CTO Achim Baumstark said, “For us, it’s clear that artificial intelligence and language models in particular will be a key competency going forward. Hence our interest in gathering experience with this technology as rapidly as possible within a clearly defined test environment.”
The Lucerne University of Applied Sciences and Arts is assisting with the experiment and will help evaluate the results.
Putting aside for a moment the sociological – not to mention ethical – implications of rendering 30% of global workforces redundant, the reality is that at present there are simply too many unknowns involved in relying on pure technology to ‘think’.
Computers can only work with data with which they are provided. If they are fed nonsense data they will produce nonsense results – what is sometimes referred to as ‘rubbish in, rubbish out’. There are also concerns that generative AI could compound existing societal biases and some leaders have called for a six-month moratorium on development to allow humans to explore the ramifications further and develop appropriate regulation.
The onus would seem to be on the developers and sponsors of these generative AI platforms to demonstrate that they are pushing an agenda that is in line with ESG principles while enhancing the plight of humanity – rather than simply pursuing profit.
Such an undertaking should make insurers feel more comfortable in embracing this new technology.
Paul McNamara
Editorial director
Asia Insurance Review