Hayford, Owen --- "Self-driving cars: Who's liable when the car is driving itself?" [2017] PrecedentAULA 21; (2017) 139 Precedent 29

---

SELF-DRIVING CARS: WHO'S LIABLE WHEN THE CAR IS DRIVING ITSELF?

By Owen Hayford

Once confined to the realms of science fiction, self-driving vehicles are now a reality. The opportunities they offer are significant. Most importantly, they could significantly reduce road accidents. They could also provide greater mobility for people who cannot obtain a licence to drive, such as the elderly, the disabled and children. They could reduce congestion on our roads by ensuring steady traffic flow, while remote parking could free up car parking spaces in urban areas for higher value uses. Finally, when we have vehicles that don't need a human driver to watch the road and be ready to intervene, we could enjoy better productivity and lifestyles because we could use the time we spend concentrating on driving to do other things.

Driver error is believed to be the main reason behind more than 90 per cent of all motor vehicle accidents.^[1] As automated driving technologies assume greater responsibility for driving tasks, the incidence of accidents due to driver error should reduce significantly. Of course, the incidence of accidents caused by deficiencies in the automated driving technologies will surely increase as these technologies assume greater responsibility for driving tasks. However, it is expected that increasingly automated vehicles will avoid more accidents than they cause.

But accidents will continue to occur, and those who suffer loss will need to be compensated. Currently, motor vehicles can operate on public roads in Australia only if they are being driven by a licensed human driver. Consequently, most motor vehicle accidents are caused by a driver’s failure to take proper care. In these cases, the 'at-fault' driver becomes liable to those who suffer injury, property damage or other loss as a result of the accident. The at-fault driver is generally insured for these liabilities by a mixture of compulsory third party (CTP) insurance (which covers the driver for his or her liability to others for personal injury or death) and comprehensive motor vehicle insurance (which covers not only the owner of the at fault vehicle for the cost of repairing or replacing that vehicle, but also for the at-fault driver's liability to third parties for damage caused to their vehicle or other property).

There are different statutory schemes in each Australian state and territory that supplement or affect the common law liability of drivers for losses arising from motor vehicle accidents, and associated insurance arrangements. The schemes are generally designed to regulate the amount of compensation that can be claimed and enable it to flow to injured parties more quickly than would otherwise occur. Some but not all of these schemes depart from the common law fault-based approach and allow injured persons to claim compensation without having to establish who is at fault. Whether fault-based or not, all of these schemes assume that motor vehicles are driven by a human driver. It is presently uncertain whether injured persons will be eligible to claim compensation under these schemes if the vehicle that caused the injury was being driven by an automated driving system, rather than a human driver.

WHO'S DOING THE DRIVING?

Driving essentially involves route finding (navigation), steering, velocity control (acceleration and braking), collision avoidance (watching the road and avoiding other road-users and obstacles), rule compliance (no speeding!) and vehicle monitoring (watching for evident vehicle faults that might affect its ability to operate safely).

Vehicles have, for many years, included technologies and systems that assist drivers with these tasks. Examples include power steering, cruise control, adaptive cruise control, anti-lock braking system (ABS) and Electronic Stability Control (ESC).

More recently, vehicles have been released that can control the vehicle's steering and velocity for extended periods of time, albeit under the supervision of a human driver who must continue to watch the road (or, to use the language of the SAE International standard for classifying automated vehicles, ‘monitor the driving environment’) and be ready to resume control of the vehicle if necessary. The SAE International standard refers to these vehicles as level 2 partially automated vehicles. The Tesla Model S is perhaps the best known example of such a vehicle.

The next generation of automated vehicles (level 3 conditionally automated) will assume primary responsibility for monitoring the driving environment in some situations (for example, on a dual carriage motorway), but on the basis that the human driver must be ready to take back control when requested. When the automated driving system (ADS) is responsible for watching the road, the human driver will be allowed to perform non-driving tasks (for example, read a book), so long as he or she is ready to take back control when requested and is alert to evident system failures.

In time, it is expected that vehicles will become capable of performing the dynamic driving task and monitoring the driving environment without the need for the human driver to remain ready to take back control. It is expected that vehicles will initially be capable of operating in this mode in only limited domains such as motorways (level 4 highly automated), but eventually vehicles will become capable of operating in this mode under all conditions (level 5 fully automated).

The diagram below summarises the different levels of automation based on the SAE International Standard J3016. The dates for each level have come from the Australia Driverless Vehicle Initiative.^[2]

As vehicles become more automated, the notion that the human driver is in control and is therefore responsible for the vehicle becomes more strained. For level 4 and 5 vehicles, it seems to make more sense to consider the automated driving system to be in control of the vehicle, and the entity responsible for the automated driving system responsible for the vehicle.

The position with respect to level 3 vehicles is presently the matter of some debate. The US Department of Transportation has suggested that the automated driving system should be deemed to be the driver of these vehicles and therefore responsible for compliance with road rules. However, our own National Transport Commission (NTC) takes a different view. It has suggested that the human driver should remain responsible for compliance with road rules when operating a vehicle in level 3 mode.

The difference of opinion has arisen because of uncertainty surrounding the requirement to be ready to take back control when requested. The vehicle will request the human driver to retake control as it approaches the end of the environment in which it can operate in level 3 mode (the operational design domain), or if there is an automated driving system failure. But how much warning will the automated driving system give the human driver? Must the human driver be ready to take back control the instant the automated driving system requests this? Or will the automated driving system give the human driver a minimum period of warning? Will the period of warning be sufficient to allow the human driver to safely resume control of the vehicle if he or she has been reading a book or engaged in other non-driving tasks? These are all good questions, which remain to be answered. Initially the answers may differ between manufacturers, as market participants work out what consumers really want.

One thing is certain, however. If the NTC approach is adopted, level 3 vehicles will be less popular than they would be under the US approach. Even if the manufacturer agrees to reimburse the human driver for all speeding fines and other infringements, the operator of the vehicle would still suffer the licence demerit points. But the NTC's approach may be welcomed by manufacturers. The US position requires manufacturers and automated driving system providers to take on greater legal responsibility, which could delay the commercial deployment of level 3 vehicles and associated safety and productivity benefits.

PRODUCT LIABILITY CLAIMS AGAINST MANUFACTURERS

As automated driving systems do more of the driving task, it is likely that accidents will increasingly be caused by deficiencies in automated driving systems, rather than the negligence of the human driver. In these circumstances, persons suffering loss as a result of an accident will not have a cause of action in negligence against the human driver, and access to compensation under statutory motor vehicle accident compensation schemes may also be difficult, for the reasons mentioned above.

Persons who suffer injury or property damage as a result of accidents caused by deficiencies in an automated driving system will need to find alternative sources of compensation.

Fortunately, Australia has well-established laws that enable claims for compensation to be brought against those involved in the manufacture, sale and use of defective products. These laws offer a proven framework for dealing with complex liability issues arising out of new technologies, and provide good reason to be confident that our existing laws will respond appropriately as the level of automation in vehicles increases.

Under these laws, a person who has suffered loss arising from a defective product may (depending on the circumstances) commence an action for compensation on the following bases:

• the common law tort of negligence (which is fault-based);

• contract; and

• breach of provisions of the Australian Consumer Law (ACL)

Often, product liability claims for defective vehicles are brought in claims of negligence or under the ACL. This is because claims in contract can only be brought by a party to the contract (that is, the purchaser of the vehicle).

Negligence

Manufacturers in Australia owe a duty of care to users to safeguard them against the foreseeable risks of injury when using the product as intended. Retailers also have a duty to guard against those dangers known to them, or those which they have reasonable grounds to expect might arise.

In the case of a motor vehicle, this duty extends not only to the purchaser but may also include others whom the manufacturer should reasonably have been aware may be harmed such as drivers, passengers and other road-users. This means that, in respect of the design and quality control of a motor vehicle, the manufacturer has a duty to build the vehicle with a degree of care appropriate to the dangers attendant upon its use, as well as to warn prospective users of its dangerous propensities.

Whether a manufacturer has exercised reasonable care will often require an examination of the state of technical and scientific knowledge at the time the vehicle was manufactured and distributed. The greater the risk of injury from the product, the greater the depth of research and testing required.

Importantly, manufacturers and suppliers of motor vehicles owe a continuing duty to purchasers and others to take reasonable care to prevent the vehicle from causing harm, including after the vehicle is sold. Failure to recall a vehicle which may cause harm may amount to negligence and give rise to the obligation to pay compensation to persons suffering loss as a result.

Contract law

Contract law is relevant because the contract under which a vehicle is sold may include express or implied warranties regarding the automated driving capabilities of the vehicle. If the vehicle's automated driving system proves to be less capable than warranted, the seller could be liable for breach of contract.

The terms of the contract of sale will also be relevant where the manufacturer or seller seeks to limit their liability through limitation of liability clauses, acknowledgements and warnings.

The Australian Consumer Law

The ACL imposes a number of statutory warranties and guarantees on the supply of consumer goods, a category in which motor vehicles are expressly included. Part 3-2 of the ACL imposes statutory guarantees on manufacturers, such that they are directly liable for:

• goods which do not correspond with their description;

• goods of unacceptable quality;

• goods unfit for a stated purpose; and

• non-compliance with express warranties.

Under Part 3-5 of the ACL, manufacturers will be held strictly liable to consumers for injury or property damage suffered as a result of a defective product. This strict liability regime removes the need to establish fault on the part of the manufacturer. Goods are considered to be defective if their safety is not of a standard that persons are generally entitled to expect. The ACL provides that in determining the standard of safety required, regard is to be given to all relevant circumstances, including:

• the manner in which, and the purposes for which, they have been marketed;

• their packaging;

• the use of any mark in relation to them;

• any instructions for, or warnings with respect to, doing, or refraining from doing, anything in relation to them;

• what might reasonably be expected to be done with or in relation to them; and

• the time when they were supplied by their manufacturer.

Importantly, there are a number of defences to a claim that goods have a safety defect. These include:

• the defect alleged did not exist when goods were supplied by the manufacturer;

• the goods were defective only because there was compliance with a mandatory standard;

• the state of scientific or technical knowledge at the time the goods were supplied was not such as to enable the defect to be discovered (the so-called ‘development risk’ or ‘state of the art’ defence); or

• in the case of the manufacturer of a component used in the product, the defect is attributable to the design of the finished product or to any markings, instructions or warnings.

Relevant considerations

The outcome of a product liability claim depends on the cause of action relied upon and the facts and circumstances of the case. Questions to be explored by the court in determining liability include:

1. The circumstances of the accident:

o who or what was in control of the vehicle at the time of the accident;

o whether the owner of the vehicle provided adequate instruction regarding the capabilities of the vehicle to the driver;

o whether any in-car instructions or warnings were deployed;

o whether the driver was adequately monitoring the situation and was sufficiently ready to take back driving responsibility;

o the interaction of the automated vehicle with its surroundings, including any lane markings or other infrastructure; and

o the actions of third parties.

2. The condition of the vehicle:

o whether the vehicle was subject to design or manufacturing defects;

o whether any unauthorised modifications had been made to the vehicle; and

o whether the vehicle had been properly maintained in accordance with the manufacturer’s instructions.

3. Representations made by the manufacturer or retailer regarding the capabilities of the vehicle, including instructions for use, markings, warnings and product advertising.

4. Whether the risk of the automated driving system failing in the manner it did were known to the manufacturer and, if so, whether the manufacturer took sufficient steps to reduce or warn of the risk.

5. Whether the manufacturer failed to adequately test the vehicle prior to its sale.

6. Whether, having regard to the state of technology at the relevant time, further testing would have revealed the problem.

7. Whether the government asked for a specific level of testing and, if so, should it have identified that the testing was inadequate.

The above list is not intended to be exhaustive, but rather demonstrates that the question of liability will necessarily be complex, and will often involve a consideration of multiple factors and parties.

THE IMPORTANCE OF ADVERTISING

With the introduction of new technology there is often a rush to emphasise the potential benefits. Unsurprisingly, there is a degree of hyperbole regarding the promise of automated vehicles, be it in respect of safety, who may operate them and what the operator may do while the vehicle is in motion.

The relevance of marketing and other statements by suppliers and manufacturers to the court’s assessment of what constitutes both ‘acceptable quality’ and ‘the extent of the safety of goods’ for the purposes of the ACL was mentioned above. It is therefore important that advertising for automated vehicles:

• is consistent with the instructions for use and safety warnings for the vehicle; and

• does not create a gap between the promise and what the technology can in fact deliver.

Advertising which depicts activities that are inconsistent with the safe operation of the vehicle could undermine the defence of a product liability claim, while advertising which creates a gap between consumer expectations and the capabilities of the vehicle could give rise to a claim that the vehicle was not of ‘acceptable quality’ and therefore in breach of the statutory guarantee.

Consumers who are disappointed when the promise of a vehicle's automated driving capabilities is not matched by the reality could also bring an action alleging misleading or deceptive conduct. Section 18(1) of the ACL provides that a person must not, in trade or commerce, engage in conduct that is misleading or deceptive or is likely to mislead or deceive. While a person may not claim compensation for personal injury or death resulting from a contravention of s18(1), there is a long history of claims brought in respect of product advertising.

In short, it will be important for manufacturers and retailers to manage the expectations of consumers as to the capabilities of automated vehicles.

CAN MANUFACTURERS LIMIT THEIR LIABILITY?

While some manufacturers have expressly stated that they will accept liability for traffic infringements or accidents caused by a failure of their automated driving system in certain circumstances, most will seek to limit their liability.

There are a number of mechanisms that manufacturers may employ to limit their liability to owners and operators of automated vehicles.

First and foremost will be the use of instructions, written warnings, acknowledgements and in-vehicle warnings to the effect that the operator of the vehicle should monitor the driving environment at all times and be ready to respond either of their own initiative or to a request by the system to intervene. The Tesla Model S owner's manual contains the following warnings:

‘Autosteer is intended for use only on highways and limited-access roads with a fully attentive driver. When using Autosteer, hold the steering wheel and be mindful of road conditions and surrounding traffic. Do not use Autosteer on city streets, in construction zones, or in areas where bicyclists or pedestrians may be present. Never depend on Autosteer to determine the appropriate driving path. Always be prepared to take immediate action. Failure to follow these instructions could cause serious property damage, injury or death.’

‘Many unforeseen circumstances can impair the operation of Autosteer. Always keep this in mind and remember that as a result, Autosteer may not steer Model S appropriately. Always drive attentively and be prepared to take immediate action.’

In addition, a dialog box appears on the user interface/dashboard every time Autosteer is activated reminding the driver to ‘Always keep your hands on the wheel’ and ‘Be prepared to take over at any time’.

The courts will not look at warnings and instructions in isolation. For example, what if the marketing material suggests that operators can take their hands off the wheel, or perform non-driving tasks such as reading a book? What if the marketing material and instructions for use include warnings, but experience shows that drivers will use the technology inconsistently with such statements or warnings? What if vigilance controls could have been included in the vehicle to ensure that the operator remains in the minimum state of cognitive awareness needed to intervene appropriately when required? A brief YouTube search turns up a number of videos of drivers using the Tesla Model S without their hands on the wheel, or their eyes on the road.

ACCESS TO INSURANCE SCHEMES AND ACTION SHOPPING

It is presently uncertain whether current compulsory third-party insurance schemes will apply to automated vehicles when being driven by an automated driving system. Some schemes require there to be ‘a person driving the vehicle’ (or words to similar effect). If the vehicle is being driven by an automated driving system, this eligibility requirement may not be met.

There is concern that persons injured by automated vehicles could end up worse off than persons involved in human-driven crashes. Although the injured person may be able to bring a product liability claim against the manufacturer of the vehicle, the process is likely to take longer and be more costly than the process available under a compulsory third-party insurance scheme.

That said, some claimants may prefer to pursue a product liability claim against the manufacturer, even though it may take longer and be more costly than claiming under a compulsory third-party insurance scheme. This is because most compulsory third-party insurance schemes cap the damages that can be recovered under the scheme, whereas the damages recoverable via a product liability claim against a manufacturer of a defective vehicle are not capped. Accordingly, the claimant may receive more compensation via a claim against the manufacturer than it could receive under a compulsory third-party insurance scheme, even if the process takes longer and is more expensive. Action-shopping by injured plaintiffs choosing a cause of action and/or defendant/insurer to maximise the quantum of compensation recoverable could become an unhealthy feature of our road trauma compensation arrangements.

To address the concern of injured persons being worse off because they cannot access a compulsory third-party insurance scheme, the NTC has recommended that state and territory governments review these schemes to identify any eligibility barriers to occupants of automated vehicles, or those involved in a crash with an automated vehicle.

Some have questioned whether compulsory third-party insurance schemes are an appropriate long-term solution to compensating persons who are injured by highly automated vehicles. Compulsory third-party insurance schemes are funded by the insurance premiums paid by owners of motor vehicles. This makes sense when it is largely the liability of vehicle owners (as drivers) that is being covered. But when we get to the point where it is the liability of vehicle manufacturers that is being largely being covered by these schemes, will it still make sense for the scheme to be funded by vehicle owners rather than by manufacturers?

Further, as vehicles become increasingly automated, it is likely that the risk of an automated vehicle causing personal injury or property damage will become quite different to the risk of a human-driven vehicle causing injury or property damage. This difference in risk profile will ultimately flow through to the cost of insuring these risks. Manufacturers of self-driving vehicles may wish to access lower-cost insurance in respect of their potential liabilities than is available under existing compulsory third-party insurance schemes.

Perhaps we may ultimately end up with new or additional statutory motor accident compensation schemes that deem manufacturers/suppliers to be liable for personal injuries or property damage caused by highly or fully automated vehicles (unless the manufacturer can demonstrate negligence by the owner or operator of the vehicle, or some by another party, such as the maintainer)? But before proceeding down this path, governments would also need to consider:

• whether there ought to be minimum levels of insurance cover for market participants in order to protect the public against being unable to hold those market participants to account (for example, due to their insolvency); and

• whether a reduced scale of damages (similar to that which applies to drivers under existing statutory motor vehicle accident compensation schemes) ought to apply.

As already mentioned, under the current system there are economic incentives for persons injured in an accident to sue manufacturers or suppliers rather than the driver, because if they succeed they will often recover higher damages than would otherwise be the case. If the requirement to establish liability under existing product liability laws was removed without a concurrent limitation on damages, this would discourage vehicle manufacturers from bringing their products to the Australian market.

CONCLUSION

As automated driving systems assume greater responsibility for the driving task, it is likely that:

• the proportion of accidents caused by a failure of the human driver to take proper care will fall; and

• the proportion of accidents caused by a defect in the vehicle's automated driving system will rise.

Those who suffer loss as a result of accidents caused by vehicle defects may not be able to obtain compensation via the same mechanisms that apply to accidents caused by driver fault. Instead, they will need to establish liability on the part of the manufacturer or others responsible for the automated driving system. Fortunately, Australia has a sophisticated product liability regime that is among the best in the world. Put simply, if a product has a defect, a consumer who suffers loss as a result of the defect can recover. Accordingly, even though there is detail still to be worked through, particularly in relation to access to existing statutory motor accident compensation schemes, there is good reason to be confident that our existing laws will respond appropriately as the level of automation in vehicles increases.

Owen Hayford is a Partner with Clayton Utz. He specialises in the transport and infrastructure sectors. PHONE [phone]. EMAIL [e-mail].

This article is based on some earlier reports published by Clayton Utz. If you'd like to learn more about the legal issues associated with automated vehicles see Driving into the Future: Regulating Driverless Vehicles in Australia (<https://www.claytonutz.com/ArticleDocuments/178/Clayton-Utz-Driving-into-the-future-regulating-driverless-vehicles-2016.pdf.aspx?Embed=Y>) and Steering the Course for Future Driverless Vehicle Regulation in Australia (<https://www.claytonutz.com/ArticleDocuments/178/Steering-the-Course-NTC-Recommendations-November-2016.pdf.aspx?Embed=Y>). The author gratefully acknowledges the contributions of others to the reports that informed this article.

---

^[1] National Highway Traffic Safety Administration (2008). National Motor Vehicle Crash Causation Survey. US Department of Transportation, Report DOT HS 811 059.

^[2] http://advi.org.au/australia/levels-of-automation/.