Driverless vehicles are moving out of science fiction onto our roads. Various manufacturers have signalled intent to make them viable by 2021 meaning that we could be driven to work automatically sooner than we all anticipated.
Whilst driverless vehicles have long been a hot topic, significant progress has been made in their implementation after the Automated and Electric Vehicles Bill 2017-2019 received Royal assent on 19 July 2018 and is now the Automated and Electric Vehicles Act 2018.
What are driverless vehicles?
Driverless cars are exactly that: cars with no driver. As opposed to a human controlling the car, passengers will be driven by an artificial intelligence which aims to get passengers safely from A to B. Manufacturers have claimed that the developing technology will make our roads safer than ever before as the biggest cause of accidents in the UK is largely attributed to human-error.
Automotive-heavyweights such as Ford, Google, Mercedes, Uber and Tesla are but a few of the names who are committed to forwarding the auto-car initiative with testing on roads having been underway since 2016-2017.
When being driven on public roads, the automated vehicles still require a person to monitor their proper orientation and “take over” when needed. This raises a significant number of questions when it comes to liability (particularly criminal liability) which will be examined in more detail later.
The technology appears to have progressed rapidly in recent years and Apple has increased the number of test vehicles to 45 as of March 2018 having started out with only 3.
To assess the progress of the A.I, it is useful to look at the average distance that the vehicles are being driven between “disengagements” (i.e. when the automated software is disabled in favour of human control). In 2017, Waymo vehicles drover over 352’000 miles and reported 63 instances where the human “supervisor” had to take control.This equates to around 5000 miles being driven before disengagement. In March 2017, it was reported that Uber vehicles required disengagement every 0.67 miles indicating the difference in success the manufacturers are experiencing.
What is clear however is that the race between these manufacturers is on as they all strive to get the finished product to market first.
How will driverless cars be used?
An obvious use for driverless vehicles is within the haulage industry and several companies are said to be testing automated trucks . “Otto”, a self-driving trucking company demonstrated their trucks on a highway before being bought by Uber in 2016. In 2017, “Embark” partnered with truck-maker Peterbilt to start testing and deploying the A.I technology in Peterbilts vehicles. Waymo is also said to be implementing the technology in trucks but as of yet, no timeline for mainstream implementation has been announced but “Starsky Robotics” became the first contender in the race to have their truck drive in fully automatic mode on a public road without a human supervisor in June 2018 so strides are being
Eventually, the A.I technology is intended to cover other kinds of transport such as busses, trains and boats as well as trucks. In 2015, the UK launched public trials of an automated pod in Milton Keynes and the City of London has been testing similar pods (effectively shuttle busses) around some parts of the city.
Potential Benefits of Driverless Cars
Safety experts predict that traffic collisions (and resulting deaths/injuries), caused by human error should be substantially reduced once the technology has been fully developed. In the USA, it has been estimated that the driverless technology could “eliminate 90% of all auto accidents….prevent up to $190 billion in damages and health-costs annually and save thousands of lives”. “TheDrive.com” (operated by Time Magazine) reported in 2017 that none of the safety experts they contacted were able to rank driverless systems as having achieved a greater level of safety than hands-on, human driving. At this time, it is difficult to determine how these safety benefits will ultimately manifest over time. Concerns have frequently been raised however in respect of the software security and how bugs, flaws or even malicious hackers may be able to affect to affect the system and compromise safety.
Driverless cars may help reduce labour costs, relieve travelers from driving and other chores associated with it. Being driven everywhere would create more time for work and leisure during journeys and would likely reduce the number of instances where a driver is distracted, texting, intoxicated or otherwise impaired. By removing the steering wheel and any driver interface from the front of the vehicle, the interiors can be more ergonomic flexibility.
It has been suggested that driverless technology could allow for higher speed limit and increased capacity on the roads. The requirement for safety gaps and stopping distances between vehicles is based on the human ability to react. With an A.I in control, it has been claimed that these requirements will no longer be needed meaning we can fit more vehicles on our roads and travelling at higher speeds (which to us, sounds like a recipe for disaster). It is also claimed that the need for traffic police will be reduced as officers will not be required to react to any poor human-drivers.
The manufacturers of the technology are also hoping that the overall cost of vehicle insurance will be reduced with lower congestion, improved flow of traffic (resulting in better fuel efficiency) and a reduced number of accidents all contributing to lower premiums.
Manually driven vehicles are reported to be used only 4-5% time, and being parked and unused for the remaining 95-96% of the time. Autonomous vehicles could (theoretically) be used continuously after it has reached its destination. This could dramatically reduce the need for parking space. When combined with the potential reduced need for road space due to improved traffic flow, this could free up huge amounts of land in urban areas, which could then be used for parks, recreational areas, buildings, among other uses; making cities more liveable.
Other related effects
By reducing the labour costs of mobility as a service, driverless cars may help reduce the number of cars that are owned by individuals as they would be replaced by a taxi/pooling shared service. This would, in turn, reduce the size of the car manufacturing market which may help improve the environment. Depending on the cost of this service, the overall household expenditure could be reduced giving people more money in their pockets. A vehicle having an “increased awareness” could help police by reporting illegal passenger behaviour but this could lead to a much-expanded mass surveillance which may not be very well received by the public.
Potential Obstacles to Driverless Technology
The above benefits may be halted by foreseeable challenges such as disputes over liability. In the event of an accident, will the passenger be liable or the manufacturer of the software?
What if the someone is killed by the vehicle but it could have been avoided had the driver disengaged the A.I? But what if said driver was intoxicated?
In terms of ethics, what if the vehicle was faced with a scenario where it could collide with either one person or a crowd of people? Should it potentially be programmed to kill one person as opposed to potentially many?
There will inevitably be a period where human drivers share the roads with this A.I and this will likely be the time when most questions/obstacles materialise. Will humans even want to forfeit their ability to drive? Should they have to?
The legal framework and government regulations may also present a huge obstacle.
Other issues that may arise would include putting people with zero driving experience in a supervisory role over an automated vehicle. What if that person has to suddenly take control?
There are also concerns about making large numbers of people currently employed as drivers unemployed, the potential for more intrusive surveillance as a result of police and intelligence agency access to large data sets generated by sensors and pattern-recognition AI (making anonymous travel difficult), and possibly insufficient understanding of verbal sounds, gestures and nonverbal cues by police, other drivers or pedestrians.
This is all before we consider the possible technical problems that that driverless cars face which are:
- Artificial Intelligence can still not function properly in chaotic inner-city environment
- A car’s A.I could be compromised, as could any communication system in place between vehicles
- The car’s sensing and navigation systems may be susceptible to different types of weather (such as snow) or deliberate interference
- The cars must be able to avoid large animals and Volvo found that software cable of recognising deer and elk was unable to recognise kangaroos
- Driverless cars will be reliant on maps that must be the highest of quality and accuracy in order to operate. What happens when a map may be out of date?
- Road infrastructure may need to be changed for driverless cars to function effective
- Companies working on the technology have faced problems with recruitment as the talent pool has not grown with demand.
Loss of Jobs
The UK currently has thousands of professional drivers who will effectively be made redundant by driverless technology within the transport industry. Professional drivers and their unions may resist its implementation as a result of job losses. Garages and repair stations may also effectively be put out of work as well as some aspects of the insurance industry as certain occupations become obsolete.
The vehicle’s location and position may be integrated into an interface which other people can access. Automative hacking is also a huge concern as well as how the data collected from trips will be used. There has also been concerns raised about terrorist attacks. What would stop a driverless car being loaded with a bomb and sent to its destination?
Whilst the environment is often cited as a benefit of this technology, there is also a chance it could have a negative impact. If people no longer have to worry about a stressful commute and can benefit from more productive time spending during the journey in addition to the savings in travel time and cost which have been oft-cited as benefits of the technology, more may choose to live out of the city. Property/land is cheaper yet if they still work within the city, the travel distance is increased along with fuel consumption which is detrimental to our carbon footprint
As mentioned previously there is certainly an ethical question of attached to this technology. What action should the car take when faced with an unavoidable crash? Does it crash into a bus, potentially killing people inside; or swerve elsewhere, potentially killing its own passengers or nearby pedestrians? How capable will the A.I be at analysing risk? Whether an automated vehicle’s capacity to correctly detect an upcoming risk, analyse the options or choose a ‘good’ option from among bad choices would be as good or better than a particular human’s may be difficult to gauge.
The Automated and Electric Vehicles Act 2018 (AEV Act)
Now that driverless cars have been defined and the pros and cons have been looked at, what exactly is the legal position?
The Act is split into two parts: the first relates to liability of insurers where an accident is caused by an automated vehicle, and the second relates to issues around public charging or refuelling points for electric vehicles.
For the purposes of this piece we will deal solely with points from Part 1 of the Act. This applies to self-driving vehicles that ‘may lawfully be put in self-drive mode on roads or other public places in Great Britain’.
The key provisions that came from the Automated and Electric Vehicles Act 2018 (The Act) is that insurers are going to be liable for damage to people and/or property where:
- An accident is caused by an automated vehicle
- Is driving as an automated vehicle; and
- Where the vehicle is insured at the time.
If the vehicle is in driverless mode when it is in a situation when it is inappropriate then the liability will fall on the individual.
In the event that the vehicle is being driven automatically, but in a situation when it is inappropriate to do so, the liability will fall on the individual and it is classed as negligence.
This was a concern during the Bill’s passage through parliament and one that defence lawyers will likely be interested to see develop: when is it ‘appropriate’ to allow the vehicle to ‘drive itself’? Without clear guidance on this specific point insurers will likely look to this section of the act for a defence to any claim whenever possible.
Would-be engineers or those who ignore software updates may also find yourself in hot water in the future. If any alterations have been made to the software which were prohibited by the policy, or if the insured person fails to install safety-critical updates then the insurance company will have an action against the individual in respect of any money paid out.
For software updates, the test to be applied is whether the insured knew or reasonably ought to have known that they were critical but what if they claim to have never received the notification? Will the onus be on the manufacturers to prove that the notification/push-message was sent?
There has been a significant level of publicity about the fact that The Act gives innocent victims a direct-right of action against the vehicle’s insurance company. The idea is to enable victims to claim compensation quickly without prolonging the process with complicated product liability claims against the manufacturers or dealing with liability disputes between insurer and manufacturer.
This direct right against the insurer will only apply where an accident is caused by an automated vehicle “driving itself” which is clarified as ‘operating in a mode which is not being controlled and does not need to be monitored by an individual’.
The Act however makes no mention to the fact that there are varying levels of automation at which the vehicle may be driven (i.e. parking only to entire journeys).
The Society of Automotive Engineers (SAE) categorised the degree of automation in 5 levels. Level 1 is where the vehicle behaves like a conventional car whilst Level 5 is when there is no input or monitoring from the “driver” required.
During Parliamentary debate however the Government confirmed that the SAE levels lacked precision and it would seek to set its own safety standard via a technical committee operating under the United National Economic Commission for Europe. Once standards are defined, it is expected that vehicles will need to pass a type-approval process to demonstrate capability to safety drive on UK roads before they can be sold (similar to how current vehicles have to pass technical standards before going to market).
It was also confirmed in Parliament that legislation will be drafted regarding the level of automation so what exactly will be the level that is deemed lawful for the Act to apply?
The Act requires the Secretary of State to publish and update a list of all motor vehicles capable of safely driving themselves and which may lawfully be used in public. Within two years of the list’s publication, the Secretary of State must prepare a report for Parliament on the impact/effectiveness of the regime and the extent to which the legislation has ensured that there is appropriate insurance in place for automated vehicles.
The Next Stage and Moving Forward with Driverless Cars
The Centre for Connected and Autonomous Vehicles (CCAV) has asked the Law Commission review the UK’s legal framework for automated vehicles. The Commission will be considering a variety legal areas. These will include (but are not limited to):
- road traffic legislation
- offences and product liability
They will also examine the use of automated vehicles as part of modern public transport networks.
The review is not intended to determine whether driverless cars will ultimately be beneficial. Its function is to consider the legal framework within which driverless technology may be implemented.
Driverless vehicles do not fit into the current framework easily. Many of the legal requirements currently in force apply only to human drivers. For example:
- How do automatic systems fit within safety assurance mechanisms such as MOTs?
- Who is accountable in an accident or crime?
- Who will decide on the safety of an automatic vehicle?
Our project will not cover drones or vehicles designed solely for use on pavements.
The review undertaken by the Commission is a three year project which started in 2018. The first year will effectively be an audit of the current law alongside an exercise to identify key issues. A paper with the results of the audit will be available before the end of 2018.
Following the review, we should have a clearer idea of how much change our legal framework needs.