Internationally, insurance regulators are focused on a few key themes, artificial intelligence (AI) and the world’s natural catastrophe protection gap among them. These were the focus in a recent panel discussion hosted by the International Association of Insurance Supervisors (IAIS).
The IAIS is a voluntary membership organization of insurance supervisors and regulators from more than 200 jurisdictions. (In Canada, the IAIS’s members include the Office of the Superintendent of Financial Institutions and the Autorité des marchés financiers.)
Artificial intelligence, the presenters say, was the group’s most requested seminar topic.
Most notably, the IAIS’s representatives told webinar participants to forget the notion that AI adoption can be achieved with a few expert hires. They also say those involved in the regulation of insurance should dispense with the notion that AI is magic.
“It’s not magic; it’s a technology. Look at it from that perspective,” says Petra Hielkema, vice chair of the IAIS Executive Committee, chair of the IAIS FinTech Forum and chair of the European Insurance and Occupational Pensions Authority (EIOPA). “What rules and what principles-based rules do we already have to supervise that?”
Instead of simply hiring a few AI experts, they say an entire (regulatory) organization needs to work with the technology to be able to supervise it effectively.
They say AI is transforming the insurance sector, creating a lot of opportunities in underwriting, pricing, claims handling, fraud management and customer engagement. “It can help reduce costs. It can contribute to financial inclusion, provide better services, quicker services to customers,” Hielkema says. “But it can’t be left unchecked. There are risks that we need to understand, risks that might be amplifying.”
Scott A. White, vice chair of the IAIS Executive Committee and commissioner of the Virginia State Corporation Commission’s Bureau of Insurance also noted that cyber risks are becoming more widespread every day, requiring discussions about data privacy and consumer protection from algorithmic bias. “These all intersect with this kind of technological innovation that kind of amplifies risks,” he says. “Not to stifle innovation, but where do we focus? How should we be monitoring the industry?”
Hielkema says there needs to be three things put in place, the first being an effective exchange of knowledge and an effort to continually build that knowledge of AI. To that end the IAIS has developed a central database of AI knowledge and research which members can query.
“Forget that you can hire a few very AI savvy new colleagues and then you’re up to speed. No. You need to make sure your entire organization starts to work with this new technology to understand it, to supervise.”
Next, she says the basics – data validation and quality – are fundamentals which need to be embraced and re-embraced, “particularly now that we see there are third-party AI models, user sets that we may not have developed in-house. The industry needs to understand how it was developed, what the data is and how it is trained. Get your basics right.”
Finally, they say there is still a lot of principles-based common sense that supervisors can have and put into practice. “If you don’t know, again, there is the database,” Hielkema adds. “There is a community of supervisors that we can also check with.” She also points out that the technology is developing quickly, but its use is also developing across the globe.
“Don’t think it’s magic. Just apply the logic of principles-based supervision. Then if you don’t know exactly what to do, reach out to colleagues because jointly, we know a lot.”
Similarly, they say colleagues are also a good place to start when trying to choose the right project. “IAIS has a good role to play by sharing those insights, by having a repository of the different things that have been done and the different techniques that have been adopted.”
Regarding protection gaps, they add that any approach needs to include risk reduction, adaptation and coordination. “Any approach to addressing protection gaps must start by embedding incentives for risk reduction and adaptation. This is the essential foundation,” says Siham Ramli, vice chair of the IAIS Executive Committee and director of communications and international relations with the Autorité de Contrôle des Assurances et de la Prévoyance Sociale Morocco.
And although they add that it is important to recognize that every country’s situation is different and not all efforts will work under all circumstances, they add that coordination is another challenge which needs to be undertaken.
“It will only work if we all act. That means individuals need to mitigate and adapt. So do cities. Insurers can provide that knowledge and provide the insurance, the products,” Hielkema says. “The challenge of coordination is to really have everyone move, instead of everyone pointing to someone else to move first. That we need to overcome. I think we can do a lot more than we do.”
