Colorado passes law to prevent AI-generated discrimination in life insurance

The news: Colorado has pioneered a law that will be effective November 14th of this year, mandating that life insurance companies establish discrimination-curbing protocols for AI use. This legislation precedes similar measures emerging across the US.

What the law says: Colorado’s leading a seismic shift in how AI is regulated within the insurance industry—setting the tone for a growing discussion. 

• The new law covers a wide range of technology used in life insurance underwriting, including algorithms, predictive models, and data collected online. 
• It aims to prevent unfair discrimination in the life insurance industry based on race, gender, and protected characteristics.
• Insurers will be required to share all components of their framework with the Colorado Division of Insurance, which will monitor for compliance with the law.

Is Colorado the first of many? Around the world, fear of the “unknown” has prompted legislators to regulate the use of AI across various industries.

• Lawmakers are increasingly focused on studying AI's role and impact on consumers.
• New York, New Jersey, and Rhode Island have already introduced bills that would regulate the use of AI in insurance settings.
• Twenty-five states, Puerto Rico, and the District of Columbia have enacted legislation affecting AI’s usage in multiple industries.

Can AI be discriminatory? AI’s only as good as the data it’s trained on. That means if the data contains bias, AI will snowball it into the results on which its users base their decisions. 

• Within the insurance realm, it’s estimated that 38% of “facts” used by AI contain bias.
• AI can also reduce human interaction, stripping empathy from customer relations.

How it impacts the insurance industry: When used responsibly, AI can help insurers process large amounts of data—speeding up and simplifying risk analysis, fraud prevention, and claims processing. However, when used irresponsibly, it can have grave consequences.

• AI-based bias can lead to disparities in access to insurance products and pricing, disadvantaging certain demographics.
• Insurance companies could turn away potential customers without reason, costing them business.

What insurers can do to prevent bias from AI: When older programs or data are used to train AI, there’s a risk of incorporating bias.

• Insurers should shift toward collecting real-time, first-hand customer data, including acquisition information, web behavior, and claims data.
• Leveraging marketing data allows insurers to phase out questionable historical data, actively correct AI biases, and provide more personalized and equitable insurance solutions.
• Removing demographic information from old datasets can worsen discrimination—avoid this tactic.

Key takeaways: AI’s not going anywhere, but neither is the fear about its impact on insurance—or our society in general. While Colorado’s the only state to pass AI laws impacting the insurance industry, it’s just a matter of time until other states, and potentially whole countries, follow suit.

Insurers outside of Colorado should get ahead of future regulation—and do their part in combating discrimination—by developing bias-reducing frameworks.