Realizing AI’s potential while addressing its risks

SMALL PHONE

PHONE

TABLET

DESKTOP

LARGE DESKTOP

XL DESKTOP

Two people meeting with coffee, papers, and a laptop

We’re harnessing AI, while dispelling the notion that algorithms using AI technology make our decisions

Authored by:
Tony Trenzeluk, Director of Government & Regulatory Affairs, Sentry Insurance
Genna Alexander, Government & Regulatory Affairs Analyst, Sentry Insurance

At Sentry, we’ve been using a form of artificial intelligence (AI) for many years—let’s call it machine learning. Our IT and data analytics teams build systems and models in house that help us better measure risk, develop and price products, and serve customers. So, we’re already realizing AI’s potential. As risk management professionals, however, we’re not blind to the technology’s potential pitfalls.

Our Government & Regulatory Affairs (GRA) team consistently observes regulators and legislators scrambling to adapt to AI’s fast-growing presence in the marketplace—especially its generative capabilities. To help them and infuse transparency in the process, our GRA team proactively provides key feedback and information to ensure regulators and legislators are well-informed about how insurers use AI, and more importantly, how we do not.

This communication with regulators and legislators helps dispel the notion that algorithms employing AI technology are making decisions that can leave the door open to discrimination, unmitigated bias, and unfair/unethical practices.

Our efforts are ongoing, as policy issues often emerge around new technology that can reshape traditional business models and consumer behavior. The rapid adoption and advancement of AI by industries, including insurance, ensures that the use and application of AI technology will continue to be a legislative topic for years to come.

Harnessing AI’s potential

Sentry uses AI models and systems in underwriting and claims management to turn unstructured historical and real-time data into valuable insights—while recognizing that conversations, expertise, and “keeping the human in the loop” remain key to conducting our mutual insurance business with integrity. The advancement of predictive analytics and real-time risk monitoring is positively transformational.

Our teams recently built an AI-based claims summary tool that generates concise narratives of claim facts. We anticipate this capability will result in considerable time savings, allowing our claims reps to spend more time with customers while we assess liability and compensability.

We also developed a predictive tool to help improve workers’ compensation claim outcomes for injured workers and employers.

Benefits resulting from the use of AI technology also include more accurate risk pricing, more efficient and cost-effective products and services, faster claims settlements, and enhanced fraud detection. Sentry views AI as one tool among many that can assist employees in the performance of their job duties and responsibilities.

Adopting regulations and guidelines

Recently, the National Association of Insurance Commissioners (NAIC) adopted its AI bulletin, outlining standards for AI implementation, governance, and risk management. More than 10 state departments of insurance have since adopted the bulletin, and four other states have created their own insurance-specific regulations or guidelines governing the use of artificial intelligence. All other states are conducting research on the impact of AI in insurance and other financial service sectors.

In Wisconsin, Sentry’s state of domicile, our GRA team worked with state and national partners to develop recommendations and proposed changes for the state’s version of the NAIC bulletin, expected to be released soon. We fostered a constructive relationship with Wisconsin Insurance Commissioner Nathan Houdek and recently held an open house where employees from various parts of our enterprise could ask him about the use of AI, his role in the NAIC, and how the Wisconsin Department of Insurance implements AI into its workflow.

At the federal level, a bipartisan Senate AI work group recently released a framework of policy priorities for AI research, development, risk management, and legislation. The framework includes recommendations for funding further innovation, testing standards, promoting competition in development, data privacy standards, and more.
Federal and state legislatures have taken significant steps to understand and govern this emerging technology, with 884 bills introduced in 2024 sessions alone that directly reference AI. Our GRA team actively monitors 37 of these bills.

Because there’s no uniform approach to how the country addresses AI, our GRA team will continue to provide regulators and legislators the feedback and information they need to make informed decisions.