The International Association of Insurance Supervisors has released a guide on artificial intelligence, focusing on consumer protection and prudential soundness.

The global standard-setting body says the application paper aims to support supervisors, insurers and intermediaries by providing guidance on good practices.

“Additionally, this paper notes that increasing application of AI can heighten the role of, and the risk from, third parties like AI model vendors,” the paper says.

“As these technologies become embedded in the sector’s operations and decision-making, the need for effective oversight to ensure their ethical, fair, trustworthy and safe use is increasingly important.”

The paper covers five topics: risk-based supervision and proportionality; governance and accountability; robustness, safety and security; transparency and explainability; and fairness, ethics and redress.

On fairness, the paper cautions that AI systems can be susceptible to biases and other stereotypes present in training and secondary data sources.

“Bias could inadvertently be programmed into AI system protocols, leading to unfair or discriminatory decisions if not properly managed.

“Furthermore, AI systems can be used to manipulate or exploit consumers’ behavioural biases, such as their willingness to pay or propensity to shop around at the renewal stage of the contract.

“While these pricing practices may exist with and without the use of AI systems, they can potentially lead to unfair or unethical outcomes if there are no adequate governance and risk management measures in place.”

See the paper here.