In 2021, Colorado Gov. Jared Polis signed SB21-169, the first of its kind, into law that establishes a framework requiring insurers to review whether the use of big data in their processes is discriminatory. The first draft of a set of rules for managing predictive algorithms and models was released in February and focuses on life insurance underwriting. Private passenger car underwriting is next, and the Colorado Division of Insurance, part of the state’s Department of Regulatory Affairs, is working on developing various rules for insurers.
Michael Conway, insurance commissioner for the state of Colorado, said at a meeting on April 6 that this work needs to be done.
“It is clear that the world of big data/AI is approaching us very quickly, in new ways, on an almost constant basis,” Conway said. “I absolutely, fundamentally believe that if we, as regulators and the industry, do not figure out how to regulate this as part of the insurance practice, someone from outside the insurance world will come in and regulate this for us, and I don’t think that will be good. for all.
“I really hope that you all will have a good and frank conversation with us, but know that we will continue to move this work forward,” he continued.
Stakeholder meetings are designed to engage industry representatives in discussion before rules are adopted. The Restriction on Insurers’ Use of External Consumer Data Act protects consumers from unfair discrimination based on race, color, national or ethnic origin, religion, gender, sexual orientation, disability, gender identity, or gender expression.
Jason Lapham, director of big data policy and artificial intelligence at Colorado’s insurance division, said that while the first segment of life insurance is still in progress, it’s time for the auto insurance industry to get ready as well.
“The first requirement for carriers is that they must check whether the use of those tools that use external consumer data and information sources lead to unfairly discriminatory results, and also establish a governance or risk management structure that is designed to mitigate the consequences for potential consumers. the harm that can result from the use of these tools,” he said.
External consumer data is defined as credit scores, social media habits, location, shopping habits, home ownership and more, Lapham added.
“One example is the use of credit histories [which] are often a reflection of historical economic discrimination against historically marginalized communities, and the use of insurance scores that include credit histories to assess risk has the real ability to replicate that discrimination by charging members of protected classes disproportionately higher premiums,” Lapham said.
Conway said the division doesn’t know what external consumer data each individual company uses, there are probably companies that use criminal history, for example.
“When we talk about testing, we are going to focus on the overarching results of any algorithm or big data engine that insurance companies use for underwriting, and then we will work backwards. see what rating factors we or the industry believe have a potentially negative, unfairly discriminatory impact. It’s a long way to go,” Conway added.
Lapham said there are potential benefits to using algorithms that include big data and predictive models, including the ability for carriers to better assess risk, which could create opportunities for insurers to cover underserved populations.
Lynn Elliot of APCIA spoke during the meeting to note that APCIA represents approximately 1,200 insurers and reinsurers in the Colorado area, which provide 57% of property and casualty insurance coverage in the state. “Senate Proposition 169 poses unique challenges for regulators and regulated companies,” she said.
Elliot also noted the efforts of APCIA and other industry companies to promote diversity, equity and inclusion.
Michael Dylon of the American Consumer Federation added: “We appreciate the insurance industry’s commitment to diversity and equity, but it doesn’t have to be limited to your workforce. refers to your business practices and the use of various factors that we talked about that he listed, including credit history, home ownership, education level, occupation, etc. And we have a lot of evidence that the use of these factors disproportionately harms blacks and Latin Americans. consumers and it unfairly harms other consumers.”
The Innovation, Cyber Security and Technology (H) Committee of the National Association of Insurance Commissioners has several groups working to define and develop a framework for the use of artificial intelligence. The Accelerated Underwriting Working Group will explore the use of external data and analytics in accelerated life insurance underwriting, which is working closely with the Big Data and Artificial Intelligence Working Group. This working group is tasked with exploring the use of big data and AI, as well as machine learning, to develop possible recommendations for managing models.
The European Union is also working on a law regarding AI. Legal initiatives include a European legal framework, a system of civil liability and a review of sectoral safety legislation, according to the EU website.