Artificial intelligence to combat underinsurance

Beyond the ethical issues it raises, artificial intelligence (AI) has changed the perception of risk in life insurance. In recent years, this technology has pushed the boundaries of risk underwriting. 

Predictive models based on AI now allow for quick issuance of life insurance policies without the policyholder having to provide bodily fluid samples or undergo medical exams. All of this is possible for increasingly higher insurance amounts, with just a few clicks from your mobile device in the comfort of your clients' homes. 

Some insurers anticipate that these technologies will also enable the insurance of individuals who cannot currently be insured. "In simplified underwriting, we began innovating with temporary insurance in 2016, for individuals aged 18 to 40. We later increased the maximum age and transitioned from term insurance to permanent insurance," said Mathieu Charest, Chief of Individual Insurance Products and Pricing at Manulife, in an interview with the Insurance Portal. 

Not more at risk 

At Canada Life, Andrea Frossard, Senior Vice President of Participating Life Insurance Solutions, has observed an expansion of accelerated underwriting programs in the industry, "especially since the pandemic in 2020." " The industry has increased the ages and limits where companies automatically order fluids (blood and urine), as part of the underwriting process," she said. According to Frossard, the predictive models introduced by most insurers have helped manage the underlying overall mortality risks and the subsequent cost of insurance. 

At Manulife, Charest asserts that the risks of simplified issue products are not necessarily greater. "We are comfortable with what we have in terms of risks for straightforward cases and have robust tools to identify higher-risk cases," he adds. 

For her part, Frossard believes that by determining when fluid samples are necessary, predictive models have accelerated the process. Insurers "will continue to order fluids for some ages and amounts, but on a selective basis rather than just broadly. The benefit of using these predictive models to assist in risk assessment is to fasttrack the underwriting process by reducing the number of clients who need to provide medical evidence," she said. 

Red flag 

Frossard points out that the predictive model doesn't make all the decisions. It's up to the underwriter to decide when medical evidence is required. "The goal is to order medical evidence in cases where we reasonably expect it will influence the underwriting decision, based on a combination of sampling methods and underwriter discretion." Statistics from the past, like smoker/non smoker, female/male, ages, could raise a red flag, she explains. 

Insuring new groups 

Charest emphasizes the importance of conducting more research to offer insurance to groups whose health conditions are currently uninsurable: "It's about seeing how medicine has evolved, developing pricing processes that follow medical advances. Nobody would have thought of insuring people with HIV 30 years ago. Now we do." 

According to Charest, this advancement occurred because individuals who are regularly monitored by a doctor and take their medications have a life expectancy similar to that of the average population. "The same goes for people with diabetes. There are also rapid advances in medicine that could change the game in the treatment of several forms of cancer," he observes. 

Charest recalls that some cancers, such as thyroid, breast, or prostate cancer, were not individually considered years ago. "Now, we classify them by category to avoid looking at all risks as one. Research results focus on each category. If, for example, we see changes in life expectancy after treatment, we will change our approach, such as offering insurance to the affected individual with a reasonable additional premium due to the improved prognosis resulting from treatment," Charest explains.