Why AI is a double-edged sword for P&C insurance

Artificial intelligence (AI) represents a double-edged sword for the Canadian property and casualty insurance industry, particularly as it relates to repetitive and rule-based tasks, speakers tell the RIMS Canada Conference in Vancouver earlier this week.

Robotic process automation (RPA) essentially looks at repetitive and rule-based tasks — a key component in the insurance world, particularly from an underwriting and claims perspective.

And while there are many benefits to using RPA in telematics, for example, it’s a double-edged sword, says Sean Hobson, senior vice president, national retail affinity leader for commercial programs at Marsh Canada.

“One of the questions in vehicle insurance is, ‘Do you drive for business?’” Hobson says during the session, Artificial Intelligence and Technology Advancements for Risk Management. “‘How far do you drive? How many kilometres do you drive a year?’

“Well, that [AI] system will tell you, ‘No, you drive actually 20 kilometres to work, not five like you told me, and you drive 50,000 kilometres a year, not 20,000 because you got a break at 20,000 and under.’

“So, it can be used against us. It can be used for us in a lot of ways.”

Telematics offers personalized recommendations and rates based on driving behaviours, “how you drive your car, how your car drives,” Hobson says. Some insurance carriers offer both discounts and surcharges based on telematics data.

Ultimately, AI will augment, enhance and streamline the industry’s roles and responsibilities. It will also assist with fraud detection and reduction.

“It’ll take that historical data, it will…join with the data that we have today and make different adjustments and help us to understand and help us to see what is actually going to happen,” Hobson says. “And in the future [it will] be predicting where there could be some fraud…And that goes throughout all [insurance product] lines.”

AI also eliminates human error, he adds.

One challenge will be integrating AI systems into business operations, says another speaker, Candace Heron, manager of claims management at Toronto Hydro. “You have silos of data; often they don’t talk to each other.”

Of course, job displacement is always a real concern. “You’re starting at a place of distrust with many people,” Heron says.

Studies estimate that up to 40% of insurance and related industry jobs could undergo major role and task changes in the next two decades.

“It’s more than just, ‘Will these AI tools take over our jobs?’” Heron says. “I think it’s a psychology that we all need to understand, that people have fear around AI.”

Like many others, she believes AI will take tasks that are very automated off people’s plates. “[AI is] able to look at data and make some analysis and conclusions faster than a person could.”

Feature image by iStock.com/Parradee Kietsirikul