The devil may be in the details, but what about insurance bias in the numbers?
For actuaries and other insurance industry workers, bias is one of the topics that Jessica Leong, CEO of Octagram, is hoping to explore as part of her contribution to the panel discussion “Competitive Intelligence – Market Trends and the Future of Insurance,” at the Women in Insurance Summit at the W Chicago City Center Hotel on May 11.
While there will be a lot of “back-and-forth” in the panel format, “in terms of latest trends, I think a lot of it is around ESG,” she said of environmental, social and governance. “On the environment side, climate change is obviously one of those big topics. Then, on the social side, there are topics around bias and assurance. This has become a big thing for society and also for our regulators.”
Around the world, from Colorado to Europe, there are “a lot of new regulations around how insurers need to test, particularly their pricing models for bias in insurance, how insurers need to govern what they were doing in terms of modelling for bias and insurance.”
This is nothing new in the insurance industry, but it is an issue that has come into sharper focus in the past few years.
“You need to actually show that you’ve tested that your models are not biased,” she said.
Even if there is no deliberate bias on the part of actuaries, for example, “there could be bias in the data,” she said. She pointed to the New York Times bestseller by Cathy O’Neil, “Weapons of Math Destruction” (2016), which looked at the social impact of algorithms and how they can be used to reinforce existing inequality.
“If you Google ‘nurse,’ you’re going to get lot of pictures of women, right?” she asked. “Google ‘computer programmer,’ you’re going to get lots of pictures of men, right? So there is just inherent bias in your data. So that could be one way bias is working its way into the system.”
Society also has to ask itself a lot of hard questions – even which questions it should ask, and how they are applied.
Are women better drivers than men? If so, should they get a better rate than men?
Or if there is a zip code that is known for having a higher crime rate, should that mean a higher insurance rate for residents of that area? “But then, that also correlates very highly with race, right?”
For Leong, “it’s an interesting blend with the math side of things, but they’re also talking about bias mathematically, but also just societal values.”
Leong said she is looking forward to being on stage in Chicago and acquired her public speaking skills through a combination of “force and through deliberate study.” She honed her skills with presentations at a previous employer, and “then worked my way up to a leadership role.” She founded her own consulting firm last year, after having served as head of data and analytics for Zurich – North America, and as the president of the Casualty Actuarial Society in 2021.
Leong said she is looking forward to taking part in this conference, and the networking opportunities it affords.
“I remember one of my first conferences after COVID. Actually, my first (post-pandemic) event was a women in insurance event,” she recalled. “I just got so much energy from that. It was amazing and I made so many new contacts. I that when so many of us are working from home like I am, you’re working remotely, when you meet in person, it’s probably even more meaningful when that happens. So I’m really looking forward to that.”