Help

BI’s Article search uses Boolean search capabilities. If you are not familiar with these principles, here are some quick tips.

To search specifically for more than one word, put the search term in quotation marks. For example, “workers compensation”. This will limit your search to that combination of words.

To search for a combination of terms, use quotations and the & symbol. For example, “hurricane” & “loss”.

Login Register Subscribe

Editorial: AI complements risk management

Reprints
Gavin Souter

Technological progress has a habit of quickly but often unobtrusively changing the way we live and work. We all know that systems and devices usually work better than they used to, but we can’t always put our finger on when or why they improved.

The benefits of some inventions, though, such as cellphones or, for those of us studying math in the 1970s, the pocket calculator, are much more immediately noticeable. Other developments are even more apparent. 

I remember being surprised and delighted in the mid-1990s when I logged on to a PC I shared with a colleague to search the internet for the first time. Calling up The New York Times’ then recently created webpage — even though I’d already read the paper copy earlier in the morning — it was clear that something had changed, and it would have far-reaching consequences for the industry I worked in and many others. 

The internet was not new, of course, but its accessibility had become much wider, and that would change our lives. 

Playing around with ChatGPT over the past few weeks, I get a somewhat similar feeling. Like the internet, artificial intelligence is not new. According to ChatGPT and good old-fashioned Google, the term was coined in the mid 1950s — although they differ on the exact date. 

But the recent public launch of the large language model by the research group OpenAI has allowed those who are interested to log in and try out AI for themselves. The result has been the rapid creation of numerous poems of questionable quality and lots of comments about the mistakes that ChatGPT makes, even though days or weeks later it corrects many of those mistakes as it continues to “learn.”

Regardless of the current shortcomings, it is clearly a powerful tool that, as it is developed and develops itself, has the potential to reshape numerous tasks and jobs.

As we report here, the insurance industry is no stranger to AI, with chatbots and other applications being used to deal with some customer inquiries, gather underwriting information and in the claims process. The more expansive capabilities of tools like ChatGPT, though, could open the sector to far more automation.

You don’t have to be a technology expert to see how it could potentially be used for tasks like generating creative ideas, streamlining complex processes and taking over mundane tasks.

There are potential downsides, too, including increased liability if AI is employed without appropriate controls, and coming up with those controls will undoubtedly involve some trial and error.

So, for the foreseeable future at least, the task of analyzing, mitigating and transferring complex commercial risks looks more likely to be enhanced by AI rather than being replaced by the technology. Unless ChatGPT knows better.