Opinion paper by Noël Ghanimé, CEO of Allianz Global Assistance France, published on Le Monde.fr, on 14 August 2016.
On Wednesday 10th August, the French Council of Ministers authorised the use of driverless cars on public roads. The rise in autonomous cars – driven by companies such as Tesla, Google and Volvo – is undoubtedly the future of the automobile industry. However, this growth raises questions in the insurance sector, not least of which is the death of the sector – ‘uberisation’, to use the buzzword.
American billionaire Warren Buffett, whose holding company Berkshire Hathaway owns Geico, recently said the following during a forum for automobile industry leaders: “The arrival of autonomous cars will not be a great moment for our insurance activity, though this will not come to pass right away.”
More than 90% of road accidents every year are due to driving errors. If autonomous vehicles make it possible to drastically reduce this figure, then we must congratulate ourselves on this progress. There will therefore be less need to insure vehicles and drivers. In a recent report, KPMG estimated that the incidence of vehicular accidents could be reduced by 80% by 2040, for a rate of about 0.009 accidents per vehicle. From this, KPMG concluded that the insurance market should therefore experience a 60% decrease by 2040.
Is the bell tolling for the insurance industry? According to Keynes, we are all dead in the long run, but decision-makers need to anticipate and manage any transitions by identifying and solving the problems that arise from these.
We can see four: the way in which the transition will unfold, the shift in responsibility in the event of an accident, new challenges regarding data, and new risks related to autonomous vehicles.
The first problem pertains to the transition from a world dominated by human drivers to one of autonomous vehicles. This shift will take place at different speeds depending on the country and on the individual. Companies will have to adapt to these differing stages, and continue to drive overall policies in a setting of very inconsistent regulations.
Sixteen American states have implemented legislation on self-driving cars. The American administration is beginning to work on vehicle-to-vehicle (V2V) communication. In other countries, self-driving cars are still forbidden, with political decision-makers not currently looking to change regulations.
The question of responsibility
The transition will also have to be managed at the consumer level. Many users of Tesla’s Autopilot have posted videos showing risky use of the system which is highly advised against by the manufacturer, such as reading the newspaper while letting the car drive along by itself, while not holding the steering wheel. At the very least, drivers and passengers will need to be trained and supported.
Another question that comes up is that of responsibility. In May 2016, the driver of a Tesla died in an accident while the car was operating in Autopilot mode. In such a case, who is responsible? Tesla considers that as the system is still in a beta phase, the driver remains the primary responsible party. Volvo feels that the responsibility should be on the manufacturer.
Google, for its part, is directing its energy towards a completely autonomous vehicle. In a case like this, insurance models will likely be decreasingly focused on individual insurance, and instead turn increasingly towards insurance for the equipment, intended for manufacturers, or for fleets of vehicles, for transport companies.
Whether or not we're aware of it, autonomous cars will be programmed based on ethical choices. For example, if an accident is inevitable, will it choose to let more damage be inflicted on the vehicle's occupants, or on passers-by and other vehicles? Who will program or personalise the algorithm? These questions will be addressed by the law, but will probably require some experimentation in order to develop long-term solutions.
New risk assessment
The third question pertains to statistics and data processing. The technological shake-up caused by autonomous vehicles will drastically transform the way in which the industry assesses risk, especially since autonomous vehicles will have a high capacity for data collection, making it possible to precisely identify the causes of an accident.
These vehicles should also acquire an unprecedented ability for assistance: they will be able to detect if a driver is tired or unwell, and whether or not he or she is able to drive, potentially give control of certain parameters to external individuals, and propose additional services. Insurance policies could also evolve over time, based on a number of parameters.
Lastly, it needs to be taken into account that connected vehicles also come with new risks. Everyone today knows about these 'flash crashes', system accidents due to the workings of algorithms.
The stock market has already seen this: an unexpected event has a major effect due to the sheep-like behaviour of a lot of software. The complexity brought on by various technologies will probably have negative effects which aren't yet totally clear to us. This is without including risks related to cybersecurity and hacking.
Decision-makers should already be looking into these four major parameters, so that the coming shake-up can be controlled and have a more effective transition.
Contact Allianz Partners
Feb 6, 2017
"Securing the future and embodying trust: insurers put to the test with the blockchain revolution" by Dan Assouline, Allianz Worldwide Partners
Securing the future and embodying trust: insurers put to the test with the blockchain revolution