Navigating the AI Revolution in Insurance: The “Approved Filings” Approach

In the insurance industry, we’ve all heard the adage: “Garbage in, garbage out.”

This principle resonates deeply in a sector where data drives everything from pricing to claims processing. But what if we could flip this concept on its head to innovate while staying compliant?

Insurance companies walk a tightrope between innovation and regulation.

A 2022 Deloitte survey revealed that nearly half of insurance executives view regulatory compliance as a major hurdle to AI adoption. This statistic isn’t just a number—it’s a call to action.

Introducing: “Approved Filings In, Approved Filings Out”

Now, let’s picture this: What if we could use our existing treasure trove of regulatory-approved documents as the foundation for AI innovation?

This isn’t just a pipe dream—it’s a potential game-changer.

The concept is elegantly simple: train AI models on data that regulators have already blessed.

The logic?

If the input is compliant, the output should follow suit. It’s like teaching a new chef using only Michelin-starred recipes.

The potential payoff?

1. Innovation at Speed: By leveraging pre-approved data, we could turbocharge AI development.
2. Regulatory Peace of Mind: Using compliant data could significantly reduce the risk of AI going rogue.
3. Resource Efficiency: Why reinvent the wheel when we have a garage full of perfectly good ones?

The fine print?

Of course, it’s not all smooth sailing. There are three major aspects we need to consider.

First and foremost is that regulations aren’t static, they evolve. As far as I can count, between 2015 and 2020, the National Association of Insurance Commissioners (NAIC) introduced more than 50 new model laws or regulations. However, this part can be automated to keep the models up-to-date.

Secondly, the bias. As an actuary, I know historical data can carry the baggage of past biases. We must be vigilant to ensure our AI doesn’t perpetuate discriminatory practices. Though, this can be addressed by training our actuaries and data scientist on AI ethics to keep them in the loop at all times.

Thirdly, and slightly the most challenging one is the Black Box Problem. Regulators might demand explanations for AI decisions, and “the computer said so” won’t cut it. However, the recent developments, and also the AI itself, can help unveil what’s really happening under the hood.

To make this approach work, insurers need to govern data like a pro. Regardless who I talk to in the industry, the first thing I tell them is to ensure they have the right data stored in the right way. Hence, treat your approved filings repository like the crown jewels.

You can automate your workflows to always obtain the latest regulatory changes and continuously feed your system to learn and course correction.

Make AI talk human by investing in explainable AI techniques and training your employees.

Last, but not least, befriend the regulators.

1. Make AI Talk Human: Invest in explainable AI techniques.
2. Befriend the Regulators: Don’t wait for them to come knocking—reach out and start the conversation.

The road ahead?

The “approved filings in, approved filings out” approach isn’t a magic wand, but it’s a promising start. It offers a path to harness the power of AI while staying on the right side of the regulatory fence.

As we stand at the crossroads of innovation and compliance, this approach could be our compass. By thoughtfully leveraging our regulatory-approved assets, we can ride the AI wave without wiping out.

The future of insurance isn’t just about crunching numbers—it’s about creatively navigating the complex interplay between cutting-edge technology and our regulatory responsibility to the public.

Are you ready to chart this new course? If not, let’s talk!