According to an analysis by the National Association of Mutual Insurance Companies, policy debates regarding the use of artificial intelligence in insurance are “basically unfounded” and “harmful to policyholders.”
The use of AI in insurance underwriting and fee production has led to concerns from some regulators, advocates, and policymakers about whether AI will lead to proxy discrimination, algorithm bias, and final changes to AI over the affordability and availability of certain sectors or classes of insurance products.
Namic said 18 states are currently discussing “defective” AI-related laws. Guidance from the National Insurance Commission (NAIC) has been added to the “ambiguous concept of algorithm bias,” Namic said.
“Contrary to what could be perceived as a deliberate social effort by regulators, policyholders are harmed by increasing efforts to enhance the concept of “equity” divorced from actuarial science,” writes NAMIC’s Vice President of Data Science Policy, AI/ (machine learning), and Cybersecurity. This leads to “an inevitable break at the heart of insurance products,” she added.
Klarkowski wrote a report aimed at dispelling five myths about the use of AI and big data in the insurance industry.
“When setting road rules, policymakers must recognize that insurance has different features and pricing than many other consumer products,” Klarkowski added in a statement. “Insurance is classified based on risk, and the insurance law requires that these risk categories be actuarially sound and not unfairly discriminatory.”
Regulations aimed at the use of AI in industry pricing must be industry-specific, and limitations on insurers’ ability to price insurers’ risks led to more availability and affordability issues, NAMIC concluded.
“The data that insurers use for risk-based pricing is actuarially sound and risk-related data, and does not include or use attributes of a particular protected class,” writes Klarkowski. “To assert that the use of data, algorithms, or AI insurers in risk-based pricing is biased or distorted, is not to represent the risk represented by policyholders that the insurance law already prohibits.”
Separately, if various impact criteria are applied to insurance, the industry’s pricing approach is no longer based on underlying insurance costs, resulting in unfairly discriminatory fees. Already, the industry is adhering to legal standards of unfair discrimination, Namic said.
topic
Insurtech Legal Data Driven Artificial Intelligence
Are you interested in AI?
Get automatic alerts for this topic.
