This year’s sprawling InsureTech Connect conference, held October 2-3 in Las Vegas, was bigger and better than ever. More than 6,000 insurance company and insuretech representatives, along with 50 insurance regulators, shared what’s happening in the insuretech space today and what industry impacts to expect in the near term. (No one predicted that the conference-ending performance by Salt-N-Pepa would be so great.) Here were some of the more interesting takeaways, soundbites and predictions, at least from our perspective:
- The use of big data and algorithms represents the biggest shift in the insurance industry in over 200 years.
- Data is already revealing secrets, letting insurers see subgroups within groups that previously appeared monolithic. Insurers are using data sets that were not traditionally available.
- Insurers are gathering “tons of data” that they believe to be highly predictive. The way the industry evaluates risk 10 years from now will make today's methods seem “medieval.”
- Body implants in the next five years will capture and transmit health status in real time. As a result, life insurers could become health advisors, coaching insureds on what they should do to live longer.
- The job of the regulator is to worry about everything. When it comes to insurers’ use of big data, regulators are most concerned about underwriting and rating.
- Among other things, regulators worry that data problems and “witches’ brew” algorithms will result in unfair discrimination. They’re also concerned that some people won't be able to get insurance because of more individualized risk analysis.
- Even accurate data can be problematic if it contains proxies for unlawful factors. In addition, “unintended bias in data is a massive problem” that needs to be addressed.
- Regulators expect senior management to have “accountability for and mastery of the algorithms” their companies are using.
- Algorithms shouldn’t be black boxes. They should be designed to be explainable.
- The sources of data used by insurers should be documented and auditable. Every factor should be predictive.
- Digital ethics are a big issue. “Anything that can’t be automated will become more valuable in the future, including ethics.”
- “Push it. Push it real good.”
Like we’ve been saying, big data and algorithms are cool, but they should be used with care. Let us know if this message starts to get annoying.