Emerging technology and the driverless car – the insurer’s perspective
[Check Against Delivery]
Thank you to PACTS for inviting me to speak here today on a subject that has generated a significant amount of interest in recent months.
This afternoon I will outline some of the key issues that the insurance industry is considering, issues that will no doubt influence the development of driverless cars.
Before I begin, I’d like to briefly outline the role of the ABI and where the insurance industry sees itself on the emergence of increasing and more sophisticated automation on cars.
About the ABI
The ABI represents the collective interests of the UK’s insurance industry.
We speak out on issues of common interest to our members and help to inform and participate in debates on public policy issues. Indeed, many of the issues that preoccupy policymakers directly concern insurers.
With human error accounting for around 90% of road accidents, the potential safety implications of autonomous technology are huge and if the number and severity of crashes decreases, so too will insurance premiums.
Industry observers will already know how much emphasis insurers have placed on technology that is proven to reduce accidents. For example, Autonomous Emergency Braking (AEB) has been proven to lower the rate of collisions that result in personal injury claims by around 20% and in anticipation of the benefits it will bring to the frequency of claims, insurers have incorporated it as a component into the Group Rating process.
Put simply, if vehicles are fitted with autonomous braking systems, they enter a lower group rating, which can significantly lower the cost for customers to insure them.
If there is a sustained reduction in claims directly attributable to autonomous vehicle technology, there will, in turn, be a reduction in insurance premiums.
And that is why we are watching the development of driverless cars with close interest.
In order for consumers to take advantage of lower premiums through safer vehicle technology, we need to ensure that the technology used is tried and tested and is understood by the driving public.
Fundamentally, the insurance industry is keen to work with vehicle manufacturers, regulators and the legal community to facilitate the adoption of this potentially life-saving technology and we are keen to overcome some of the issues.
Key issues for insurers
At this stage in the development of the technology, there are two principal issues to carefully consider:
- A potential shift in the transfer of liability in the event of an accident
- Challenges and opportunities around increasing amounts of high quality data
1. Transfer of liability
The issue of liability invariably gets raised in policy discussions regarding autonomous vehicles. This is eminently sensible. When autonomous vehicles become involved in accidents, resolving the question of fault will require considering novel and in some cases challenging questions that have not been asked before.
The situation at present is clear: even with autonomous technology features on cars, liability rests with the driver.
And, in line with the development of increased automation within aviation, it is likely that the driver will continue to be held liable in the event of a crash if they are able to step in and intervene, overriding the technology by making control inputs themselves.
As with pilots, effective system monitoring for drivers will become a skill and a key safeguard in the prevention of accidents, requiring modifications to the way people learn to drive.
The key change – and the potential shift to product liability – comes when the driver is not expected to oversee or monitor the vehicle and when they have ceded full driving responsibility to the car itself.
To make the picture clearer, it is worth arriving at a common definition of the differing levels of vehicle autonomy. For today’s presentation, I will use the definitions as set out by the National Highway Traffic Safety Administration (NHTSA).
As we have said in our response to the DfT’s consultation on the testing of driverless cars, our initial view is that if a system fails on a fully autonomous vehicle causing the vehicle to crash, liability would rest with the vehicle manufacturer or system manufacturer.
However, this would only be the case when the driver had ceded full control to the vehicle and was not ‘monitoring’ the systems and could have had the ability to intervene, take over, and prevent the crash.
The potential shift in liability is likely to occur when a driver has actively given complete control to the vehicle and as a result has no option intervene, or when utilising a system with no controls at all, such as a google pod car.
The legal precedents already established relating to product liability will provide manufacturers of autonomous vehicle technology with a very strong set of incentives to make their products as safe as possible.
In the overwhelming majority of cases, they will succeed. However, despite these efforts, there will inevitably be some accidents attributable in whole or in part to defects in future vehicle automation systems.
It is possible that manufacturers will be brought in to legal proceedings following an accident involving an autonomous vehicle by either the insured driver or someone they have injured. Whether or not there is a complete shift in liability from driver to vehicle is likely to depend on whether there is a clear option for human intervention.
While the position is unclear, and as the technology develops, insurers may consider continuing to settle claims on behalf of the insured driver but then seek to reclaim the damages paid from manufacturers. While this will raise complex new liability questions, there is no reason to expect that the legal system will be unable to resolve them.
The question is whether the legal and regulatory frameworks will keep pace with technological developments.
2. Driverless cars and data
For the purposes of monitoring its performance, it is highly likely that the adoption of driverless technology will go hand in hand with the increased uptake of more advanced telematics technology which will form a key part of the vehicle’s systems.
These systems will become increasingly sophisticated, able to remotely transmit far more detailed information about a vehicle’s activity than is presently available, which will enable insurers to gather previously unavailable information on telemetry.
The proliferation of data that will result from this creates a whole host of both challenges and opportunities for insurers.
Fighting fraud
As you will know, the motor insurance industry is particularly afflicted by fraud with some people attempting to defraud insurers by making false or exaggerated claims, or by lying about who they are or their driving history.
As a result, insurance fraud is a big problem, not just for the insurers but for the police and for everyone who has legitimate car insurance as honest customers have to bear the brunt of fraud through higher premiums.
New vehicle technology provides a welcome opportunity to challenge this.
The potential benefits of driverless technology are more obvious for avoiding – or mitigating – collisions but will the technology deter and help eradicate induced or staged claims? Possibly not.
With the vehicle recording every direction simultaneously and measuring the forces involved in a collision, the data collected will allow investigators to determine the exact cause of an accident when it occurs and this will be extremely helpful in disproving many fraudulent claims. With on board technology monitoring every inch of a vehicle’s movements, dishonest claimants will no longer be able to exaggerate or lie about the cause of an accident.
Hacking and cyber threats
A key data challenge is the threat associated with the deliberate misuse of a car’s data systems.
Autonomous systems will rely heavily on internet connectivity making them intrinsically vulnerable to cyber manipulation. Hackers may be able to override a car’s system to re-route it toward a particular destination, or overwhelm it with high volumes of internet traffic.
And by intercepting and tampering with mobile communications and vehicle software updates, cyber-criminals could transmit malicious code or, in the worst case scenario, send new and dangerous instructions to the vehicle’s software systems.
The task for everyone with an interest in promoting the benefits of driverless cars is to find and close the vulnerabilities now before the technology is integrated extensively into mass market vehicles.
The importance of data sharing
With increasing use of more sophisticated event data recorders (black boxes), vehicle manufacturers will be gathering an increasing amount of information.
To maximise the consumer benefit from the increased use of in-vehicle technology, it is essential that the underlying in-vehicle systems are based on standardised, interoperable, secure and open-access platforms.
This is not necessarily to profile the risk of the driver – indeed, the onset of driverless cars may refocus the risk the driver presents to an insurer to a focus on the risk of the car itself.
The users of driverless cars should be able to consent to access by third party services who may offer a wide range of ancillary services. For insurers specifically, these services might include more sophisticated claims notification platforms, allowing the rapid processing of a claim; including the quick recovery of the vehicle and provision of a replacement vehicle. But this can only happen if the data collected by a vehicle can be shared and so to maximises the potential that is offered by the new technology – and in so doing improve consumer outcomes – data must be shared.
In addition, by showing the performance of autonomous systems at a macro level, shared data may also influence the development of the legal and regulatory framework governing driverless technology. Indeed, if the safety implications of autonomous cars turn out to be nearly as high as some have predicted, this may lead to automated systems being mandated or also heavily incentivised.
Conclusion
Looking to the immediate future in terms of the testing of driverless cars, we do not envisage any significant issues that will restrict the availability of suitable insurance products.
During the testing phase, it is likely that vehicle manufacturers may choose to self-insure, paying any damage claims themselves as they may have a better understanding of the risks posed by autonomous vehicles.
The insurance industry will use the data gathered via the testing of driverless cars to build up an accurate picture of the risk they pose and the earliest insurance products for use by the public will be modelled on the testing of driverless cars. With the development of this technology, insurers may base the premium paid less on the personal profile of the policyholder and more on the specification of the car as it likely that the car might have a larger bearing upon the risk posed than the driver.
Beyond this, there are still a large number of unanswered questions when it comes to how industry and government will approach the issue of driverless cars. Today, I have sketched out a number of possible scenarios that are themselves dependent on a wide variety of factors. As one stone is unturned so another issue emerges.
However, I would strongly argue that we have good cause to be optimistic. The potential benefits of this technology are huge, and none of the challenges that I have spoken about today are by any means insurmountable.
Ultimately, although the future looks bright for the driverless car technology, it may take years to incorporate driverless cars into our everyday lives and so, it may be a while until we see the main effects on insurers.
The insurers who are able to make use of the increasing volumes of data - building policies to recognise the shifts of risk - are likely to move ahead of their competitors and should find opportunities to make a profit in what is – and what will remain – a very competitive market.
Thank you.