May 26, 2020

The multitude of personal data going hand in hand with the explosion of the market for connected objects gives insurers an increasingly detailed view of their customers’ behavior. But collecting sensitive data can be troublesome and raises a question: how can we regain control of our data at a time when risk is extremely personalized?

The current world health crisis has been accompanied by a new surge in the monitoring of health and geolocation data, seen as one of the means to curb the pandemic. Mobile applications have blossomed as a means to “track” coronavirus in several of the affected countries. In China, citizens were given a colored bar code (green, yellow or red) to indicate what they were entitled to do or not do, ranging from a ban on the use of the metro to quarantine. South Korea, meanwhile, has developed an application that locates the most affected areas. In France, Orange has provided its subscribers’ (anonymous and aggregated) geolocation data to follow the population’s movements during the epidemic, and the government’s tracking application, StopCovid (still at the project stage) has been debated for several weeks. StopCovid is able to follow and identify people who have been in contact with an infected person, and could be “dangerous for our civil liberties”, according to La Quadrature du Net. Meanwhile, Covidom, an application that remotely controls patients who are carrying or suspected of carrying Covid-19, has been in operation since March 9th.

Although today these data tracking systems are involved in an extraordinary crisis, they are not new. Our behavior, from the number of steps we take every day to the flexibility of our repeated actions, is continuously being captured and analyzed by cookies on our browsers and applications, but also by our voice assistants, connected watches and other intelligent trackers. The home automation market is booming, carried notably by the growth of voice assistants, which are becoming the new standard. According to forecasts by the Juniper Research firm, the number of home assistants should reach 8 billion worldwide by 2023 (compared to 2.5 billion at the end of 2018). The health sector, meanwhile, can count on the “quantified self” wave, a trend towards measuring everything in our everyday performances, from jogging to the amount of oxygen in our blood.

These data give insurance companies the opportunity to get closer to their customers, enabling them to follow the development of risks in real time and to offer prices based on customers’ profiles and habits. The trend also appeals to the insured, according to a study conducted by Accenture in 2016, which reveals that 80% of the interviewees “are looking for personalized advice, offers, prices and communications, whether for car, home or life insurance policies”. Almost as many – 77% – said they were “happy to provide data linked to their usage and behavior in exchange for a lower premium”.


Prevention is better than cure

Insurance premium bonuses for the most virtuous customers are already common in the United States and Great Britain. Certain insurance companies reward customers who agree to wear a connected bracelet to prove that they do regular physical exercise. The facial analysis software, Chronos, developed by Lapetus Solutions, enables insurers to calculate life insurance premiums based on a short questionnaire and a selfie, linked to a variety of indicators (such as body mass, pathologies, smoking, etc.) to estimate life expectancy. The behavior of policyholders is no longer based just on their past history but on their everyday life, and even their future.

Prevention is increasingly part of the everyday, and requires no effort. The CareOs connected mirror at the Cardif Lab’ answers a voice command, and provides access “with just one look” to all the information necessary to assess and improve your routines, play with your reflection and learn more about your health: skin examination, acuteness of vision tests, virtual make-up trials, personalized healthcare, etc. Collecting and analyzing data in healthcare and well-being increases individuals’ knowledge and control over their bodies. For the elderly, these systems also afford greater autonomy. Taking preventive measures and thereby promoting greater well-being at home as safely as possible on the health front is the goal set by Cardif France’s Génération Care, part of the start-up Life Plus, a company committed to healthy aging and the prevention of fragility among the elderly and dependent. Génération Care enables these groups to measure their own health (pulse, blood pressure, weight, physical activity, etc.) at home and provides remote follow-up by a medical team.

For Renaud Dumora, CEO of BNP Paribas Cardif “developing prevention is a ‘win-win’ for an insurance policy,” as “greater risk prevention leads to a reduction in the social cost of damages as well as better-adapted protection”. However, this must be done in an ethical way: the extreme personalization of insurance policies must not harm insurability, he points out, evoking the risk of “excessive demutualization.” On the contrary, it must be able to increase insurability: “The wealth of data, notably in medical databases, could enable insurers to develop solutions for certain at-risk populations.”


Intelligent cars and homes

The growth of “behavior-related insurance” is not restricted to the health sector. “Pay as you drive” policies came into existence in the car insurance sector some fifteen years ago, with a similar principle: offering preferential prices for the best drivers and penalizing others, using an electronic system installed in the vehicle that assesses a driver’s aggressiveness.

The same is true for home insurance policies: in Italy, the H@bitat Homebox home insurance launched by BNP Paribas Cardif relies on a home automation box and smoke, flood and breakdown sensors. If there is a problem, the customer receives a warning by SMS, email or telephone. The insurer has thereby recorded a 20% drop in accidents by comparison with a conventional product. The future should be promising for this type of insurance, given the attraction of connected homes in France whether in terms of comfort and safety, or the reduction of energy consumption.


The question of tracking policyholders

The collection and use of this often-sensitive data nonetheless raises doubts. According to a BCG survey, 40% of French and American people claim to be concerned by the way health insurance companies manage their data. Notably in health areas, the data that is harvested can be beyond the users’ control. A discussion on an instant messaging service or a search engine query on a given pathology, regular junk food purchases and too few steps taken every day also provide indications of an individual’s health, if all these data are combined. It’s possible that in the future, this type of data could also be taken into account to determine the price of your health insurance.

This scenario is hardly likely in the short term in France, where the Evin Law bans health insurance companies from offering “behavior-linked” pricing structures. The National Centre for IT and Liberties (CNIL) also oversees the collection and processing of personal data, notably on health. Data protection regulations prevent these data from being transferred or sold, except with explicit consent from the user, or for research purposes, as in the case of the Health Data Hub, a medical database set up by the French government. Meanwhile, “pay as you drive” car insurance policies are authorized in France on certain conditions: a “connected vehicles and personal data” compliance pack published by the CNIL in October 2017 stipulates that data relating to trip records are indeed personal data and must be protected as such.


Regaining control over our data

To reassure customers, insurance companies rely on securing their data and the advantages of digitalization. Blockchain-based intelligent contracts enable insurance policyholders in certain Nordic countries to be compensated automatically and in total security in the case of unemployment. Other systems give customers the control: in the car insurance sector, “a data protection approach from the outset is preferred. This can be reflected by introducing dashboards that are easy to configure, providing the user control over the data,” say Dominique Paret and Hassina Rebaine in their book “Véhicules autonomes et connectés” (Autonomous and connected vehicles) (Dunod, 2019).

The same approach is recommended by Mathieu Cunche, a professor at INSA-Lyon and member of the INRIA Privatics team, with the introduction of secure boxes at people’s homes to which connected objects within the home can be connected in order to avoid any interaction with an outside server. “With this system, data are still collected, but they are not sent to Fitbit servers (Ed.: acquired by Google) or Apple, for example,” he explains, evoking ownCloud and Cozy cloud. “This increases citizens’ level of control over their own data, as it decides who can access which types of data and with what level of accuracy.” In this way, once can knowingly classify the level of personalization of the risk in a future insurance offer.