Insurance Leaders Urge Exclusion of GLMs from AI Oversight
Insurance industry leaders are urging state regulators to exclude generalized linear models (GLMs) from new oversight plans targeting artificial intelligence (AI) systems.
GLMs are a staple in property and casualty insurance pricing, and the National Association of Insurance Commissioners (NAIC) recently developed an AI Systems Evaluation Tool to assess AI applications in the industry. This tool, discussed in an October 8 meeting, is slated for completion in November. Industry voices, like Kristin Abbott from the American Property Casualty Insurance Association (APCIA), argue that GLMs are not AI and should remain separate from the tool's scope. Abbott pointed out that GLMs are well-established, transparent, and traditional—contrasting sharply with the dynamic nature of AI models.
Furthermore, Lindsey Stephani from the National Association of Mutual Insurance Companies (NAMIC) voiced concerns that including GLMs might dilute the tool's focus. By concentrating on new AI technologies, regulators can better detect novel risks, keeping the integrity of the current actuarial systems intact. Consumer advocates and health insurance representatives are also weighing in. They advocate for clearer definitions and examples within the tool to streamline regulatory compliance efforts. Notably, Randi Greene Chapman from the Blue Cross Blue Shield Association called for NAIC committees to review the tool further, ensuring it is both rigorous and appropriately targeted.
Key Aspects of the AI Systems Evaluation Tool
- Focus: Evaluating novel AI technologies for risk detection.
- Feedback: Hosted an October 8 meeting for stakeholder input.
- Finalization: The tool is expected to be finalized in November.
- Exclusion Request: Industry leaders advocate for excluding GLMs from review.
- Regulatory Review: Suggested review by NAIC committees on financial conditions and market regulation.
The debate continues as insurance professionals await the finalized version of the tool, which promises to shape the way AI systems are integrated and monitored across the industry. By focusing on these emerging technologies, the industry aims to manage novel risks while maintaining the stability of established processes like those using GLMs.