The phrase “Big Brother” reminds many of us of a popular reality television show where members of a household are constantly being watched and recorded. I suspect we rarely remember this phrase in connection with a work of fiction titled 1984, which was written by George Orwell in 1949. In the dystopian society depicted in the novel, citizens are observed and influenced through the government’s (aka “Big Brother’s”) use of media and technology tools. Today, most people have an adverse reaction to the notion of being watched by the government, yet they readily consent to the sharing of personal information for the “right” to post details of their lives on social media networks for all to see.
Technology is mediating our lives, and we are becoming more disconnected from each other as we become increasingly connected to the world around us through our “Black Mirror”1 devices. They connect us to our families, work and the world at large, causing our reality to mirror Orwell’s fictional world. Heaven forbid these devices are misplaced, broken or stolen, disrupting what has become a basic need for both digital immigrants and digital natives alike.
Media psychology is an emerging discipline that studies the differences between how digital immigrants and digital natives respond to media and technology. It is becoming important for insurers to recognize and understand these differences as the nature of the engagement between insurers and consumers changes from one mediated by agents to one mediated by technology. We are living in the age of the mobile app through our black mirror devices, requiring less human contact in our daily transactions. Future sales of life insurance will be to the next generations who want their insurance on-demand, inexpensive, customized and deliverable through a mobile app.
Traditional life insurance policy issuance and underwriting reflects none of these attributes. Depending on age, gender, smoking status, policy size, state of health and pre-existing conditions, the insurer incurs significant costs to cover tests on bodily fluids, attending physician statements, medical testing and examinations, cognitive testing, and analysis of nonmedical data such as financial data, motor vehicle records and credit reporting data. The collection, examination and assessment of all this information take time and still may lead to an adverse decision for the applicant or insurance at prohibitive rates. The latest innovation in insurance technology, often referred to as InsurTech, is rapidly removing the obstacles that make acquiring insurance a lengthy and cost-prohibitive process for those with certain conditions and unresponsive to consumer demands for customization.
InsurTech firms employ external big data to data traditionally collected by insurers to develop risk profiles more reflective of individual lifestyles and behaviors. Statistically predictive algorithms are proving just as effective at classifying risks as traditional methods at a fraction of the cost and time. The result is more responsiveness to consumer demands for an increasingly accelerated, customized and budget-friendly experience.
Machine learning algorithms are part and parcel of actuarial analysis tools for increasing the understanding and management of insurance risks. Credit data has long been used in property and casualty insurance to understand the propensity that an insured is likely to have a claim. It is now being used in life insurance predictive modeling to understand lifestyle behaviors of insureds and assess the credibility of health information that is self-reported on insurance applications. Credit scores and other predictive variables aid in the determination of which medical tests are necessary to facilitate risk classification. Fewer medical tests results in immediate cost savings to the insurer, and those cost savings can be passed to the consumers demanding them.
This issue of The Actuary is a monograph on how using big data and machine learning algorithms can transform the governance, marketing, underwriting, issuance, analysis and management of insurance. Most of the new tools in the actuary’s toolkit are open source and in need of a model governance framework (Alahakone and Andrews). Using the tools of market segmentation (Diede) is the first step toward better understanding the complex needs of consumers and the best deployment of analytics to gain the greatest competitive advantage (Vohra and Hutchinson). The acquisition, quality and strategic use of data (Paris) is the foundation for machine learning models. It drives the results and informs decision-making through the application of statistically reliable algorithms and actuarial judgment (Larson, Leemhuis and Niemerg, and Granieri, Heck and Tafoya). Improving data generating processes using distributed ledger technology (DLT) is the next frontier insurers will need to settle (Carruthers, Bai and Shirra). DLT can improve the reliability of data used at every level of the insurance organization, especially data used in predictive machine learning algorithms.
The future of insurance is clearly predictive. There are many pathways actuaries can pursue to acquire data science skills. Actuaries are best positioned to become data scientists for the insurance industry, given the depth and breadth of their insurance subject-matter expertise. The combination of expertise, mathematical and statistical aptitude, and computer hacking skills is the ubiquitous definition of a data scientist. The SOA has developed a certificate program designed to help actuaries complete their transition to become one.
Please use this issue of The Actuary to guide your journey to become a data scientist and help steer your company to a prosperous future state of health in an ever-changing, technology-mediated world.
- 1. Cellphones, tablets, virtual reality, augmented reality and computers; Black Mirror is a British science-fiction anthology television series that examines modern society, particularly with regard to the unanticipated consequences of new technologies. It can be viewed on Netflix. ↩