The mining of social media and the use of big data (vast amounts of available internet data that can be analysed and exploited) by insurers to analyse consumer behaviour will change the face of insurance. The basic principles of insurance include the good faith sharing of information between insurer and insured, a pooling of good risks with bad risks, and assessing a fair premium for policyholders. Paradoxically regulations, that require insurers to price more competitively and estimate their reserves more precisely to see that they are sufficiently capitalised, are leading insurers to find ways to price insured risks more accurately.
It is possible using computer algorithms to tap data from social media posts and other internet sources to evaluate the risks of insuring a particular person. The information that is posted and an individual’s “likes” can help assess personality traits. It is claimed that people who use superlatives like “always” or “never” imply overconfidence and are bad driving risks. What you like and don’t like, how you write and how many exclamation marks you use could be used against you in assessing your motor premium. It is possible by mining big data to find out all sorts of information about what people eat, what they do, how they drive and what they buy from the pharmacy. Devices such as sensors in vehicles and health-monitoring watches and step-counters will reveal data relevant to motor, life and disability risks. What you post about yourself on social media could in future determine what premiums you pay for insurance or whether you can get insurance at all. If you have a genetic test to predict possible future illness, that information should be disclosed to your insurers when you take out a new policy and could be used to discriminate against individuals whether you disclose the information or not. If you don’t disclose the information, your policy could be cancelled from inception because of the failure to make the disclosure. If you do, you may pay higher premiums or struggle to find insurance.
The Equality Act does not label discrimination as unfair if it is based on objectively determinable criteria “intrinsic to the activity concerned”. The question is going to be asked whether discrimination between policyholders on prohibited grounds, such as disability or health status based on the analysis of social media and big data to find out intimate details about each individual, is based on evidence intrinsic to the activity of insurance. Insurers, in terms of their duty of good faith, will have to reveal to a potential insured the basis of any discrimination which could lead to challenges against the use of internet sources under equality legislation. In 2014 insurers in the UK agreed not to take account of predictive genetic tests when deciding on insurance cover unless it is to the insured’s advantage and undertook to be transparent when they do use that data. Some compact is going to be necessary between insurers and regulators.
A balance will have to be found in order to keep the principles of insurance alive without any unfair discrimination. If a balance is not found, compulsory pooling of risks may be the result in relation to certain kinds of insurance which will undermine the social value of insurance and its role in the economy.
Likes and dislikes is the stuff of discrimination, so we need to find ways of balancing the ways that we deal with other people, not the least in the world of insurance.
*Patrick Bracher is from Norton Rose Fulbright South Africa