An IBM study has found most insurance industry leaders believe generative AI is essential to keep pace with competitors. However, only a quarter of customers want anything to do with the technology.
The study is based on a survey of 1,000 insurance c-suite executives and 4,700 insurance customers. CEOs in the survey were evenly decided on whether generative AI was a risk versus an opportunity although 77 percent who responded said generative AI was necessary to compete.
Customers were not so keen: just 26 percent said they trusted an AI to dispense accurate and reliable advice.
Unsurprisingly, IBM – which has an array of AI products that it would be delighted to sell to customers – reckons the solution includes implementing a strongly governed, ethical AI to deal with trust issues and "use AI to connect the underlying risk data and address long-standing insurer and financial service provider technical debt."
Or perhaps Big Blue could simply listen to customers, only 29 percent of whom are comfortable with generative AI agents providing service, according to IBM's figures.
Per the study, investments in generative AI are expected to surge by more than 300 percent between 2023 and 2025 as organizations move from technology pilots to implementation. IBM's rivals, including Microsoft, have bet big on such a move, although analysts and traders have hinted at the AI bubble potentially popping in the not-too-distant future.
Another study from Salesforce showed data and security worries were also holding back enterprises, with only 11 percent of surveyed CIOs saying the technology had been fully implemented.
- AI chatbot gets green light to hallucinate your investment portfolio
- Microsoft Bing Copilot accuses reporter of crimes he covered
- AI chatbots amplify creation of false memories, boffins reckon – or do they?
- From RAGs to riches: A practical guide to making your local AI chatbot smarter
Even when implemented, the pay-off from AI projects can be far less than hoped for by overexcited executives. It all depends on how the technology is used. As an example, some large enterprises have taken a step back from products such as Microsoft Copilot over security and governance concerns – not with Copilot itself, but thanks to a realization that access and permissions need to be configured correctly before allowing Copilot to crawl over an enterprise's data.
The findings of IBM's survey and the conclusions drawn indicate a worrying tone-deafness in the AI industry. Customers don't want to deal with generative AI agents. Rather than try to solve the problem with, er, more AI, it might instead be better to consider where generative AI is appropriate and where it is not. ®