January 12, 2018

UK begins journey to autonomous future

The UK is to begin its journey to an autonomous driving future in 2018, with the enactment of the Automated and Electric Vehicles Bill. The Bill, which is based on insurance industry proposals, will extend compulsory motor insurance to include the use of autonomous vehicles.

The road to regulation in the UK is being watched closely by those in Europe and further afield to gauge how it may work in practice. This Bill is a suitable attempt at mirroring the present insurance position, which it is hoped will help kick-start this emerging technology and further demonstrates the government's commitment to ensuring the UK is at the forefront of this developing technology.

Indeed with Uber recently contracting with Volvo to purchase 24,000 Volvo XC90 to form a fleet of driverless cars between 2019 and 2021 such regulation is needed to set the groundwork as demand for this technology accelerates. On the basis an autonomous vehicle is not limited to shift work as a human driver, those cars alone could replace every non-Uber taxi and cab in London. These developments underline how the technology could rapidly change the face of transport, manufacturing and the economy in the next few years.

However, this is very much the first piece in the puzzle, with further regulation anticipated as the technology develops. Domestic and European legislation will likely need to be amended to allow for autonomous vehicles to be tested without a human driver in the vehicle. This significant shift in regulation is needed to bring the UK and Europe into line with the legal position in US states such as Florida and Michigan, where autonomous vehicles can operate without a human presence.

The Bill maintains a single insurer model, where a motor insurer covers both the driver’s use of the vehicle and the autonomous vehicle technology, whilst allowing insurers a right of recovery from the manufacturer where the vehicle is at fault, pursuant to established product liability laws. Accordingly current "state of art" defences will be preserved for manufacturers. 

While the Bill retains quick-and-easy access to compensation for claimants, the position will be more complex where an accident involves a semi-autonomous vehicle.  If a driver fails to take back control of the vehicle in an emergency situation or where the autonomous function fails to operate correctly, determining the appropriate apportionment of liability will be difficult. However there is no reason to think the courts will be unable to accurately determine fault.

Autonomous vehicle users and their insurers need an equally efficient system of recovery from manufacturers if the system is to function. The supply chain for automated vehicles will be necessarily complex, involving both hardware and software manufacturers, including software developers, software providers and telecoms service providers. Each company in the chain and their insurers will need to be aware of their potential exposure to claims. We expect insurers, manufacturers and other stakeholders to agree a recovery process as soon as possible.

As data will be key to understanding the risks inherent in these vehicles from both an underwriting and claims perspective, further regulation concerning what data should be provided to insurers both prior to and following an accident is likely. This is likely to form the new battleground given that autonomous vehicles will create a veritable gold mine of personal data for both private companies and users, which may well be commercially sensitive.

This issue is already beginning to be addressed internationally. Global data protection watchdogs recently called for autonomous vehicle users to be given control over who accesses the data generated by their vehicles. The resolution passed by the watchdog advocated privacy controls that allows users to withhold access to different categories of data generated by these vehicles. It remains to be seen whether international consensus will be secured.

This issue is also being addressed at EU level, given the interplay with 'personal information' under the impending General Data Protection Regulation. European insurers have called for regulators to ensure motorists, not vehicle manufacturers, control personal data generated by their vehicles. Ensuring all stakeholders are on an equal footing will allow data to be accessed so policyholders' premiums can be calculated by insurers. This will be a necessary requirement if the current insurance model is to function.  Utilising Council of Bureaux organisations across Europe as a third party clearing house for data could be a useful development that may satisfy stakeholders.

In order for a fully workable insurance system for autonomous vehicles to be implemented, a clearer definition of autonomous vehicle is required to ensure regulators and insurers are able to classify and insure vehicles appropriately. We expect the Bill to be tightened to limit the definition to fully autonomous vehicles that do not require any driver input. If this does not occur, drivers may be prone to place too much reliance on non-autonomous vehicles, resulting in increased crash frequencies.

Motor manufacturers are currently bringing to market new vehicle models with increasingly sophisticated advanced driver assistance systems (ADAS). These developments are anticipated to reduce crash frequency and severity with resultant reductions in premiums. However conditional automation, where a vehicle can operate autonomously under certain conditions but hands back control where the driving task is outside the vehicles scope, could lead to over reliance by some drivers.   

Driver education will be vital in the next 12 months as vehicles take to the roads with advanced driver assistance systems. Should drivers place too much reliance on non-autonomous vehicles we may see increases in crash frequency. Such vehicles are not autonomous and it is incumbent on manufacturers to use appropriate marketing terms for conditional automation systems to ensure consumers are not being misled.

For users of ADAS systems, standard driving and managing conditional automation require different skills; with automation-management being very much a learned skill. In the short term, this is perhaps the greatest hazard risk managers face in harnessing the advantages of this developing technology. This may well make the road to automation for insurers and risk managers a little bumpy.