When back in 1903 Michigan Savings Bank’s president was advising Henry Ford’s lawyer Horace Rackham not to invest in the Ford Motor Co. for the reason ‘The horse is here to stay but the automobile is only a novelty — a fad’; he was probably clueless that he would be long remembered in history for his erroneous perception about the future.

The technology is now world apart from where it was in 1903 and where it is today. Since the construction of Cugnot’s steam-powered tricycle in 1769 the automobile has changed from an engineering to a social commodity. The modern world is now its doors to welcoming automobile vehicles sprinting out and about on public roads with no or little input from a traditional driver-in-charge.

In November 2017 the then Chancellor promised to have fully driverless cars on the UK’s road in use by 2021. As the matters were before Covid-19 pandemic, the UK seems perfectly on its course to rolling out driverless cars on public roads by 2021. The driverless or fully automated cars are now a reality and one should expect experiencing them on our roads in near future.


What is a true automated or driverless vehicle?

It seems pretty obvious that a driverless vehicle implies a vehicle without a driver our eyes; however, have yet to see an automobile on a public road without a human driver controlling it. It is even harder to imagine the causes of road traffic accidents involving such highly sophisticated motor vehicles and envisage the issue of liability in such incidents which are yet to occur.

In order to recognise the yet invisible issue first we should understand where does today’s technology stand and how automobile industry is embracing this technology.

Generally speaking a robotic or self-driving car, also mistakenly known as an autonomous vehicle, connected and autonomous vehicle, automated car, or driverless car is a motor vehicle that is capable of sensing its environment and moving safely with little or no human input.

While the above terms are often interchangeably used, they characterise a degree of technical variation which makes one vehicle slightly different to the other. We are already familiar with cars on road having some self-driving capabilities, such as, cruise control, parking assist, adaptive headlights or lane-keep assist functions etc. The car manufacturers are continuously adding these features to more models each year. Therefore, it becomes very relevant to have understanding of the difference between the various terminologies used in respect of highly or fully automated vehicles and how the law implies in each case.

As a starting point a driverless car is self-driving, but a self-driving car isn’t necessarily driverless. Likewise, most self-driving cars are automated but they are not autonomous.

SAE International (previously known as Society of Automotive Engineers) has created a ‘Levels of Driving Automation’ standard that defines the six levels of driving automation, from no automation to full automation. This classification divides automated vehicles into 6 categories, each with an increasing amount of automation thus, decreasing human intervention.  The list outlining the specifics of each level is as:

  • In SAE Level 0, the car has no self-driving capabilities at all. The human driver does all tasks related to operating the vehicle. Modest vehicles manufactured in 80’s and 90’s fall into this category.
  • In SAE Level 1, an automated system on the vehicle can sometimes assist the human driver.  The driver has overall control of the vehicle with assistance of a single aspect of automation, such as, steering, speed or brake control at one given time. No more than one of such aspect can simultaneously be used in conjunction with the other.
  • In SAE Level 2, an automated system on the vehicle can actually conduct some parts of the driving tasks while the human driver adheres to the driving environment and performs the required driving tasks as needed. Here the automation falls short of selfdriving because a human sits in the driver’s seat and more often is required to take control of the car in many driving situations. It requires full driver attention who must be ready to take over at any time. These exist today in the form of systems like advanced cruise control, parking assist, lane keep assist, and automatic braking. More than one components of driving assist can be used alongside.
  • In SAE Level 3, vehicles contain the lowest-tier system that is classified as an automated driving system. An automated system can actually conduct some parts of a driving task and also monitor the driving environment in some instances, but the human driver must be ready to take back control when the system requests.  In a number of situations the vehicles can make informed decisions for themselves, for example, accelerating past a slow-moving vehicle but―they still require human override. The driver must remain alert and ready to take control if the system is unable to execute the task. These exist today in vehicles for use in motorway environments where the lanes of the road are clearly marked.
  • In SAE Level 4, an automated system can both conduct the task of driving and monitoring the environment. It does not require human interaction in most circumstances however; a human still has the ‘option’ to manually override. Operation of this system is limited to certain environments and conditions though. In some circumstances such very bad weather conditions, the vehicle might not allow the driver to engage self-driving mode. These systems are currently being tested by Google, Uber, Apple, and Samsung.
  • In SAE Level 5, the vehicle controls all driving tasks under all conditions that a human driver could perform. Level 5 cars won’t even have steering wheels or acceleration/braking pedals.

Based on these policy definitions, a true driverless is Level 5 while self-driving is Level 4 or 3. An autonomous vehicle at levels 4 and 5 is although self-driving, but a self-driving vehicle at level 3 is not autonomous. An automated car does not have the level of intelligence or independence that an autonomous car has. Therefore, driverless and autonomous are nearer to synonyms, as are self-driving and automated.


Accidents caused by automated or driverless cars and the question of liability

Whilst the technology is yet to extensively develop and highly autonomous vehicles are initially going to share public roads with vehicles without automation, accidents involving automated or autonomous vehicles seems inevitable, so, complex liability disputes are round the corner. The main issues are likely to involve arguments around system failure, driver fault and owner’s failure to properly maintain the vehicle. There have already been some accidents in the US involving highly automated vehicles and at least one fatality has been recorded during the trials of autonomous vehicles in addition to a test driver of automated vehicle being disqualified from driving in the UK.

In addition to direct collisions with vehicles, there are some other challenges that may arise from the introduction of highly or fully automated vehicles causing disruption, for instance, traffic being blocked if a self-driving car freezes when confronting unexpected weather conditions or unknown obstacles (including, possibly, leaves or plastic bags) and issues such as increased congestion if many self-driving vehicles are introduced before private or human-driver car use has reduced.

The current law in many parts of the world relating to traffic rules for vehicles is either governed by the Vienna Convention on Road Traffic 1968 (Vienna Convention) or its predecessor the Geneva Convention on Road Traffic 1949. Both treaties share the concept of a driver being in charge of vehicles and that the driver must at all times be able to control their vehicles. UK signed the treaty in 1968 but has not yet ratified it.

Pursuant to the current Road Traffic Act 1988, all drivers have a duty to other road users to drive with due care and attention.  If this duty of care is breached and a person is injured, the victim may be able to claim compensation from the vehicle driver and this is when insurers step into claim process.

Human error is a very prominent factor behind most road traffic accidents. Some common causes of accidents include failing to look properly, misjudging other road users’ movements and being distracted, drunk, tired, careless or in too much of a hurry.

The questions, however, arise if the damage, including personal injury, is caused by an automated or autonomous vehicle – for example – whose liability it will be and how the victims would be compensated? Is it fair that a person should be able to claim compensation from a vehicle driver or in-charge when that person wasn’t actually driving the vehicle, was not in control of it, or did not breach their duty of care, or in contrast, a machine can be held negligent?


Legislation on accidents caused by the use of an automated vehicle

The UK has already legislated in the area of civil liability for when things go wrong with the way autonomous vehicles operate with the Automated and Electric Vehicle Act 2018.

The aim of the Act is to provide a quick and smooth path to compensation for death, personal injury or property damage caused by an automated vehicle by keeping the compensation route within the motor insurance settlement framework, rather than through a product liability framework.

In situations where the vehicle occupant is driving the vehicle in meanings of conventional driving, then the normal rules of negligence would apply i.e. if a human driver fails to exercise reasonable care in avoiding an accident, he would be liable. When the automated driving system is engaged, civil liability on the ‘driver’ would be replaced by the new legal provisions.

The first part of the Act clarifies that insurers are required to deal with all claims when the vehicle is ‘driving itself’ i.e. operating in autonomous technology mode. The insurer of the vehicle would be directly liable for any damage caused.

Section 3 of the Act confirms that incidents involving automated vehicles will be subject to the usual rules of contributory negligence where the accident or damage was to any extent caused by the injured party. Logic suggests the apportionment of liability in contributory negligence matter will depend upon the level of automation in play in each particular vehicle. However, the insurer or owner of an automated vehicle is not liable to the person in charge of the vehicle where the accident that it caused was wholly due to the person’s negligence in allowing the vehicle to begin driving itself when it was not appropriate to do so. In addition, insurers are able to exclude or limit their liability for damage caused by an insured individual making ‘software alterations’ that they are prohibited from carrying out under their insurance policy, or failing to install software updates that they know (or ought reasonably to know) are safety-critical.

The insurers also have a right to recover costs from the manufacturer particularly, when a technology failure causes an accident.

Finally, the owner will be directly liable if the vehicle has not been insured pursuant to the Road Traffic Act 1988.


So far so good? Perhaps, not!

The government’s efforts on supporting the development of autonomous vehicles and protecting the victims rights coincide with the future of mobility however, the legislation is still in development phase and therefore at present it fails to address a number of key issues.

The Automated and Electric Vehicle Act 2018 is currently limited to fully-automated vehicles only. The principles set in the Act are quite broad and there are several terms used in the body of the Act which have not been defined.

The Act clarifies the term ‘driving itself’ simply as ‘operating in a mode which is not being controlled and does not need to be monitored by an individual’ but it makes no mention of the varying levels of automation a vehicle can have as listed above. For this reason, it offers no guidance as to which level would be deemed lawful or safe for the purposes of automated mode. Moreover it provides no concrete guidance about when would and would not be an appropriate time to let an automated vehicle drive itself.

The term ‘accident’ is not defined anywhere within the Act.

The Act also leaves some issues open, such as the way in which the insurer can reclaim from manufacturers and whether there are any defences available?

The Act is also silent about the need for obtaining a driving license to operate an autonomous vehicle and any criterion which would be required to test the driver/operator’s suitability, qualification and competency with regards to safe driving and installing/updating software updates or ability to interact with automated system. A new digital highway code may need to be introduced and a bespoke licensing regime may need to be established.

Although Association of British Insurers has welcomed the government plans yet it is has warned that the drivers must not be given unrealistic expectations for the foreseeable future as the insurers are not expecting autonomous cars to have sufficient back-up features to allow drivers to completely disengage from the road. The insurers representing body has previously taken the view of the driver of an automated vehicle as the pilot of an aircraft who is expected to be held liable in the event of a crash if he was able to step in and intervene, overriding the technology by making control inputs himself.

What isn’t included in the Act is any legal duty on manufacturers, such as ensuring safety-critical software is implemented resting the onus of responsibility for installing safety-critical software with the human user. Therefore, in claim process the insurers may want to make investigations into the potential exclusions in liability before agreeing to compensate. There may also be challenges in determining whether a driver should have taken control of a highly autonomous vehicle or not.

Liability disputed matters would have to be resolved by the courts in cases that may well require complex technical evidence and involve the manufacturer evidence. Access to black box information from such vehicles will be crucial as well as complicated due to privacy concerns and it could form an area of litigation.

Therefore, guidance is needed as to the potential loopholes for owners and drivers regarding the required software updates, and when, where and how it is and is not appropriate to allow a self-driving vehicle to drive itself. More thought needs to be given to ensure proper training is in place for the driver or operator when driving the vehicle, or when they have legitimately handed over partial or full control to a semi-automated or fully autonomous vehicle. Questions around liability such as was the car being properly monitored when an accident occurred, was the human driver’s reaction appropriate in taking control at a given time and how long is an appropriate reaction time when in a semi-autonomous mode need to be explained?

Furthermore, human-automation interface isn’t the only problem that Compensators will have to solve in the future. The laws need to be fully standardized and infrastructure needs to be supportive of automation.



Limited trials of driverless cars on the UK’s public roads are already underway with the Government’s aim being for them to be operational for public use by 2021. Connected and autonomous vehicles (CAVs) are expected to become a reality on UK roads by 2021/22. However, privately owned fully autonomous, driverless cars are rather far off in the future. It would be some time before a fully autonomous vehicle becomes commonplace nonetheless the technology is going to completely change the way people get around.

Automated driving has the potential to initiate a road traffic revolution, but clear standards for how the technology is going to be used and what back-up systems will be in place to protect road users against system failure are yet to be set.

While the term “driverless” is often used to describe these technologies, the reality is that entirely removing the need for a driver (and therefore automating steering and other controls) is a longer term goal for most vehicle types. For now, they will be ‘automated’ cars and not ‘driverless’ and even in fully automated mode, drivers will still need to be fully trained and sober.

A review of the legal framework is currently being conducted by the Centre for Connected and Automated Vehicles (CCAV), with assistance from the Law Commission of England and Wales and Scottish Law Commission. Development and implementation of relevant regulations are anticipated via a number of statutory instruments within the next few years.

Horseless carriages propelled by gasoline might attain speeds of 14 or even 20 miles per hour. The menace to our people of vehicles of this type hurtling through our streets and along our roads and poisoning the atmosphere would call for prompt legislative action even if the military and economic implications were not so overwhelming… [a quote from the US Congressional Record in 1875]


  • Share this article!

Leave a Reply

Your e-mail address will not be published. Required fields are marked *