The Financial Services Regulatory Authority of Ontario (FSRA) has published the Fair Treatment of Consumers in Uses of Big Data Analytics in Auto Insurance, a report from the regulator’s auto insurance data and analytics strategy technical advisory committee.
FSRA says the report’s findings will inform how it sets and delivers on its priorities related to auto insurance rate reform and underwriting regulation. “We also hope this report will help the industry implement big data analytics (BDA) solutions in the auto insurance system,” they write.
“The report would offer an opportunity to build trust and mutual understanding between consumers and the industry, particularly in areas of high complexity such as BDA,” they add.
The advisory committee, formed in November 2020, has a mandate to provide advice to FSRA on the consumer impacts and regulatory implications of the use of artificial intelligence (AI) and BDA in auto insurance – critical issues in consumer protection, they say. They report, they add, explores how to treat consumers fairly in response to these issues. “It also attempts to address how data is collected, treated, processed, communicated, and utilized to inform decision making by all stakeholders in the Ontario auto insurance market.”
Themes explored in the report include model fairness, including rating model key elements, how data is treated and interpreted, and model bias. It discusses transparency and recommends the creation of a consumer survey regarding the industry’s disclosure practices. Support for vulnerable consumers and how consumer vulnerabilities could be considered and accommodated in model-related decisions are also discussed.
The committee’s recommendations include further defining vulnerable consumers and communities and work to develop mechanisms for monitoring the treatment of vulnerable consumers. “FSRA should consider vulnerability in the underwriting and rate regulation reform,” they add.
“BDA refers to the use of algorithms and advanced analytics capabilities to make or inform decisions, based on patterns, trends and linkages of new data sources,” they write, adding that BDA is used across the auto insurance value chain in Canada. While it allows companies to improve risk prediction, serve customers and develop new products, they add that the benefits come with costs and risks when not managed properly. Concerns expressed in the report include those about new risks emerging from data-driven automation – undetected data issues causing unintended harm or reinforcing existing societal disadvantages, along with the opacity of BDA models and increased individualization of insurance leading to affordability issues among high-risk consumers.
“FSRA seeks to understand current and potential BDA practices in the property and casualty (P&C) insurance market. FSRA’s aim is to ensure that consumers benefit from these new practices while also being treated fairly,” they write.
“FSRA has adopted the Canadian Council of Insurance Regulators’ (CCIR) and the Canadian Insurance Services Regulatory Organization’s (CISRO) guidance, Conduct of Insurance Business and Fair Treatment of Customers, which generally addresses the fair treatment of consumers. This report, however, focuses specifically on the fairness of BDA applications.”