Driverless cars create new insurance risks

By Michael Majstorovic*

The emergence of autonomous vehicle (AV) technologies unlocks vast benefits for the community through greater safety, mobility and productivity.

However, AVs, or driverless cars, will have profound implications for many sectors, including insurance.

AVs create new risks for insurers to consider as they examine the impacts on motor accident injury insurance (MAII) schemes, product liability insurance and cyber insurance.

For MAII schemes, understanding the technological components and different levels of automation involved in AVs will be essential for insurers to effectively manage risks and assess claims. If MAII schemes become redundant, insurance will still be needed. Product liability will be required, given the potential for technical failures, and cyber risks can threaten the computerised and connected nature of AVs. However, with a currently limited understanding of AV technologies, some of those risks may be volatile and unquantifiable.

MAII schemes to change

Australian states and territories have their own CTP schemes that compensate people injured by motor vehicles. Essentially, the schemes indemnify drivers and vehicle owners for third party injuries for which they may be liable. Some CTP jurisdictions operate on a no-fault basis, while others are hybrid or at-fault schemes where the injured person must establish fault on the insured owner/driver of a vehicle.

Australia also has national injury insurance schemes (NIIS) that provide compensation on a nationwide no-fault basis for catastrophic injuries from motor accidents.

There are three ways in which MAII schemes can change.

Risk allocation – new understandings of accident risk

When considering accident risk, insurance premiums are based on personal metrics, like the driver’s age, residence, vehicle and past records. With the emergence of AVs, underwriters will need to develop understandings of newer complex risks about technology in the vehicle. That might involve examining the age, reliability, accuracy and potential tampering of software and hardware technologies. For example, assessing how an AV performs in hazardous weather conditions, road closures and unforeseeable events. It is important these factors are considered as hardware costs can be expensive and result in significant losses.

For purposes of causation in fault-based and hybrid jurisdictions, assessments must be understood pursuant to the different stages of automation under SAE Standard J-3016, which has been endorsed by Australia.

Lower automated (levels 0-2) AVs, which already are on Australian roads (eg, lane-keep, self-parking), are likely to be unproblematic as the driver remains in significant control and attentiveness over the vehicle.

Higher levels of automation (levels 4-5) are also unlikely to be problematic, as the automated driving system controls all aspects of the dynamic driving task with little or no human assistance. Liability may not lie with a human driver there, unless the human has failed to maintain the vehicle (eg, software updates, tyre changes). However, the law is not overly adequate in enabling a shift of liability to an AV.

Issues are more likely with conditionally automated (level 3) vehicles, where the system executes the dynamic driving task but requires the driver to resume control or respond appropriately when necessary.

As the Law Council of Australia outlines, issues might arise where the system has failed to detect a possible collision (eg, due to software defect) and a human driver intervenes to avoid collision but, in doing so, acts negligently in a way that may contribute to a crash.

Causation issues will become more prevalent as individuals inevitably put more trust in AVs to perform the driving task. Eg, in North Carolina, a semi-autonomous Tesla crashed into a police vehicle while the driver was watching a movie. Causation issues can lead to higher and unquantifiable risks for insurers which they may be reluctant to cover.

Necessity of MAII schemes – AV safety expectations

Human error contributes to more than 90% of motor vehicle accidents, due to susceptibilities like aggression, distraction and fatigue. It is projected that automated driving systems, with their ability to make expeditious and effective decisions without human involvement, will significantly reduce road trauma.

Eg, within two years of its introduction, Tesla’s Autosteer feature saw the AV crash rate per million miles reduce from 1.3 to 0.8. Scholars have argued that private car ownership will decrease, with the expectation of widespread, cost-efficient, automated ridesharing services. Therefore, AV technology will inevitably lead to reduced CTP premiums and, some have argued, make CTP schemes altogether redundant.

However, as strong empirical data on higher AVs on public roads are yet to be seen, and surveys indicate many individuals are opposed to owning driverless vehicles, it is unclear whether CTP will become completely redundant over coming decades.

For motor-related insurance outside MAII schemes, like voluntary comprehensive coverage for property damage, it is unlikely AVs’ expected safety standards will affect those policies. Theft, vandalism and natural disasters are inevitable and AV owners may still wish to protect their vehicles from those perils.  

Shift to product liability insurance

Although MAII schemes might become less essential, there will be an increased need for other lines of insurance, like product liability. That means the onus for effecting insurance will shift from drivers to manufacturers.

Product liability coverage will become necessary because accidents can still occur but will likely be caused by technical failures. Eg, while it would be rare for a human to do something blatant in contravention of road safety, like drive on the wrong side of the road, it is possible an AV could. That might be due to a product design error or defect (eg, a faulty sensor).

As with CTP insurance, legislative reform might be necessary to ensure minimum levels of insurance are purchased by manufacturers to protect the public from potential insolvency. South Australia has enacted legislation that requires AV manufacturers to have approved insurance policies in place before they can trial AVs on public roads.

The need for cyber coverage

Cyber risk is generally an unfamiliar consideration in the motor insurance industry. Dependence on the computerised and connected nature of AVs creates significant threats of data breaches, GPS spoofing, privacy violations and cyber extortion.

There is also the potential for malicious interference with a vehicle’s critical safety components, like brakes. That raises concerns for large-scale cyber-terrorism, as hacking a single manufacturer could make all their AVs vulnerable. Cyber threats can involve significant losses and might lead to uninsurable and unquantifiable risks for insurers.

The unpredictability of cyber events means insurers and reinsurers will need to carefully assess whether manufacturers and software developers have incorporated adequate security and mitigation strategies.

*Michael Majstorovic is the University of Technology Sydney winner of an AILA student prize for achieving top marks in insurance law subjects. This article is an edited version of the first part of a research essay on autonomous vehicles. The second part, which examines the inadequacy of the current framework in clarifying liability when someone is directly injured from an automated driving system, will be published in the June issue of Resolve.