View source for Definition:Unfair discrimination

⚖️ '''Unfair discrimination''' in insurance refers to the practice of treating similarly situated [[Definition:Policyholder | policyholders]] or applicants differently in [[Definition:Rating | rating]], [[Definition:Underwriting | underwriting]], or [[Definition:Claim | claims]] handling when the difference is not justified by actuarially sound risk distinctions. Insurance inherently involves discrimination in a technical sense — [[Definition:Underwriter | underwriters]] differentiate among risks every day — but the law draws a sharp line between legitimate [[Definition:Risk classification | risk classification]] and unfair treatment based on factors unrelated to expected [[Definition:Loss | loss]]. Charging two drivers with identical records and vehicles different [[Definition:Premium | premiums]] solely because of race, religion, or another protected characteristic is a textbook violation.

⚙️ State [[Definition:Insurance regulation | insurance regulators]] enforce unfair discrimination prohibitions through rate review processes, market conduct examinations, and consumer complaint investigations. The standards vary by jurisdiction: some states explicitly ban the use of specific factors such as [[Definition:Credit score | credit-based insurance scores]] or gender in certain [[Definition:Line of business | lines]], while others permit these variables as long as they are actuarially justified. When a carrier files its [[Definition:Rate filing | rates]] for approval, regulators evaluate whether the proposed classification plan produces results that are not unfairly discriminatory — meaning every rating distinction must tie back to demonstrable differences in risk. Violations can result in fines, mandatory rate adjustments, and reputational damage that erodes market trust.

🔍 The rise of [[Definition:Artificial intelligence (AI) | AI]] and [[Definition:Predictive analytics | predictive analytics]] has intensified scrutiny around this issue. Algorithms trained on historical data may inadvertently encode biases, producing [[Definition:Rating | rating]] outcomes that serve as proxies for protected characteristics — a phenomenon regulators are increasingly alert to. Several states and the [[Definition:National Association of Insurance Commissioners (NAIC) | NAIC]] have launched initiatives to develop frameworks for testing algorithmic fairness in insurance pricing and [[Definition:Claims management | claims]] decisions. For [[Definition:Insurtech | insurtech]] companies building data-driven products, demonstrating that their models do not engage in unfair discrimination is both a regulatory requirement and a market differentiator, especially as consumer advocacy groups and legislators push for greater transparency in how insurance decisions are made.

'''Related concepts'''
{{Div col|colwidth=20em}}
* [[Definition:Risk classification]]
* [[Definition:Rate filing]]
* [[Definition:Insurance regulation]]
* [[Definition:Predictive analytics]]
* [[Definition:Actuarial justification]]
* [[Definition:Market conduct examination]]
{{Div col end}}