Download as pdf or txt
Download as pdf or txt
You are on page 1of 23

Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249

https://doi.org/10.1007/s13198-020-00978-9

ORIGINAL ARTICLE

Review and analysis of the importance of autonomous vehicles


liability: a systematic literature review
Mohamed Alawadhi1 • Jumah Almazrouie2 • Mohammed Kamil3 •

Khalil Abdelrazek Khalil1,3

Received: 17 December 2019 / Revised: 30 March 2020 / Published online: 17 April 2020
Ó The Society for Reliability Engineering, Quality and Operations Management (SREQOM), India and The Division of Operation and
Maintenance, Lulea University of Technology, Sweden 2020

Abstract The introduction of autonomous vehicles (AVs) liability. Further research needs to be done on how users
has led to a shift in liability from driver to AV. This paper perceive liability and how liability evolves with the con-
is a systematic review of the literature on liability with siderable changes in the law.
AVs. Articles were selected from EBSCO, SCOPUS and
PROQUEST and then sorted based on inclusion and Keywords Autonomous vehicles  Self-driving vehicles 
exclusion criteria. The findings reveal that most such arti- Driverless vehicles  Liability
cles have been published in law journals and transport
journals. The articles mainly discuss the steps being taken
by developed countries, such as the US, the UK and Ger- 1 Introduction
many; the modification of existing laws; and the formula-
tion of new laws. The articles also address the shift in A recent trend in the automobile industry is the develop-
liability from humans to AVs and affirm that there is no ment of autonomous vehicles (AVs) in preparation for the
general rule on liability for AVs. Researchers conclude that introduction and mass implementation of driverless vehi-
liability would depend on the particular details of a situa- cles. Conventional transport would then be replaced by
tion, the role of the driver, the level of autonomy exercised automated transport—giving full control to a driverless
by the vehicle, and the environmental factors among other vehicle. Shladover (2018) described the basic concept of
considerations. Earlier papers have also suggested that AVs as partially or fully replacing human intervention with
ethical considerations should be studied when dealing with electronic or mechanical devices, thus making the transport
driverless. With the advancement of technology and the
pace at which the automobile industry is evolving, the topic
& Mohamed Alawadhi of AVs has become of great interest to researchers, engi-
U16101432@sharjah.ac.ae neers, government bodies and industry experts in recent
Jumah Almazrouie years.
Jumah.almazrouie@daep.ae Freemark et al. (2019) study explored how AV tech-
Mohammed Kamil nology affects cities and found that cities have made only
mmohammed@sharjah.ac.ae limited preparations, with bureaucrats concerned about the
Khalil Abdelrazek Khalil potential risks of AVs, such as increased number of kilo-
kabdelmawgoud@sharjah.ac.ae meters travelled per vehicle. There is thus an opportunity
1
Engineering Management Department, College of
for upgrades in city design so it can be compatible with
Engineering, University of Sharjah, Sharjah, UAE AVs (Duarte and Ratti 2018). AVs could increase sprawl
2
Reliability Assurance Department, Dubai Aviation
because travelers would be willing to commute farther
Engineering Projects, Dubai, UAE (Hamidi et al. 2015), which can increase energy con-
3 sumption and segregation among travelers based on their
Mechanical and Nuclear Engineering Department, University
of Sharjah, Sharjah, UAE income (Johnson 2001). The interaction between different

123
1228 Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249

modes of travel depends on choices made in the design of autonomous technology. Technological progress con-
infrastructure, such as the amount of space allocated for tributes to the development of new transport systems to
streets (Schreurs and Steuwer 2015; Riehl 2018). AVs can ensure accessibility and progressive and sustainable cities.
reduce parking needs, for example, because they can New transport systems are needed, especially in major
increase shared travel, and they can park themselves else- cities where there is high-density traffic. Constructing more
where. This could make land now utilized for parking roads is not possible in already developed cities, and
available for other purposes. However, these positive and constructing roads around a city can negatively impact the
negative effects of AVs are conjectural. environment.
The uncertainty concerning autonomous vehicles Marçal et al. (2017) affirmed that mobility is one of the
therefore it is in the context of extensive experimentation. key issues of modern society discussed all over the world.
Automobile companies test AVs on public roads all around AVs are consequently expected to be an important part of
the world even though the complete schedule for unveiling the new transportation system and have hence become an
AVs in the market is still unknown (Freemark et al. 2019). area of interest for automotive research and development
If people continue to use conventional vehicles along with projects and the most popular research topic in modern
AVs, minor changes in transportation systems may be history (Fan and Chang 2016). AVs will affect not only
made. On the other hand, transportation systems need to urban mobility but also the transport system and the soci-
change completely when the whole traffic system becomes ety. Many issues related to the development of AVs need to
automated. City policies will change based on evaluations be addressed—for example, the effects of autonomous
of the level of public interest in AVs (Niles 2019). driving on mobility behavior, the interaction of humans
Accordingly, government officials will have an important with machines and other issues related to safety (Schreurs
role to play in setting the agenda and formulating policies and Steuwer 2015). The introduction of AVs will change
regulating AVs (Marsden and Reardon 2017). city structures (Zakharenko 2016), although it is still
In their research, Consilvio et al. (2019a, b) stated that unclear how this evolution in traffic systems will affect
population increases lead to increased transport needs, people’s lives.
which in turn cause more problems related to mobility, AVs are expected to change transportation modes,
including pollution, accidents and congestion, especially in whether directly or indirectly. According to Nikitas et al.
cities where population density is high. There is a strong (2019), along with the transport revolution prompted by
need for governments to plan effective transport networks AVs, other developments will occur simultaneously in
to address mobility problems and handle traffic move- public transport, such as driverless metros in major cities in
ments. The current problem is determining the optimal the Western world (Powell et al. 2016). Similarly, Attias
configuration of an existing road network in terms of fac- (2017) argued that the transport revolution will continue in
tors such as directions, capacities and signals (Porter 2018). the direction of AVs, with other technical objects, such as
Additionally, Consilvio et al. (2019a, b) predicted that buses, also becoming autonomous in the near future.
private and public transport will be automated in the future. Sustainability is one of the key aspects to be considered
Automation will greatly impact city infrastructure, requir- when designing AVs and developing infrastructure
ing a complete redesign of land use regulations (Madadi (Mladenovic and McPherson 2016). According to Suresh
et al. 2020). The current situation is a transition from and Manivannan (2014), autonomous driving consumes
manual to autonomous driving, with increased interactions less energy, which not only reduces pollution but also
between the two and between AVs and infrastructure. offers other benefits, such as reduced accidents, reduced
According to Vleugel and Bal’s (2018) research, cities injuries and deaths, less traffic congestion and more
occupy more people than the past, and also urban popu- opportunities for people who cannot drive.
lation will grow with time, increasing transport needs. Evaluating the societal impact of AVs, Schoitsch (2016)
Infrastructure development is costly, but it is necessary to found that AVs can reduce resource consumption, increase
accommodate more cars. Large cities require huge invest- mobility opportunities for older persons and people with
ments in infrastructure development, while small cities are special needs, reduce parking needs and decrease envi-
unable to manage large investments. For a city to be sus- ronmental effects. Safety is another goal of AVs, and safety
tainable, the supply and consumption of energy and goods risks and privacy violations have yet to be addressed. For
should be economical. This is not the case a present in most example, it is not yet clear what disruptions AVs can cause
cities. in society, given that they involve people handing over
Vehicle manufacturing is undergoing huge advances in tasks and decisions to machines.
technology, with AVs being the most noteworthy advance. Kyriakidis et al. (2015) claimed that AVs have gained
Although AVs have not entered the mass market yet, the attention of stakeholders because of their ability to
conventional cars already feature some elements of transform society and personal mobility. With maturity,

123
Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249 1229

AVs are expected to change the everyday lives of indi- partially. Likewise, if the occupant does not utilize an
viduals and transform traditional transportation standards. override mechanism to control the AV in an unforeseen
To achieve the desired transformation, this evolution in situation, liability is assigned to the occupant (Sparrow and
transportation should bridge manufacturing and technology Howard 2017). In most cases, however, human occupants
to deliver smart solutions and create an environment that do not incur liability while driving AVs.
welcomes the adoption of AVs. In conventional driving, drivers are liable in any colli-
Urooj et al. (2018) investigated legal liability in acci- sion (Gurney 2013), but this is not the case for AVs, and
dents involving privately owned AVs and other vehicles, this difference has significant implications. One the main
pedestrians or property. Current laws put the liability on benefits of AVs is that they can reduce accident rates. In
the human driver, since the actions of the driver cause the contrast to accidents involving conventional vehicles, in
collision, which is not the case with AVs. Similarly, laws accidents involving AVs, the automobile manufacturer, the
governing computers hold the operators liable for any software company or a third party is liable (Lederman et al.
abuse, misuse or noncompliance, but these laws do not 2016), since there is a chance that the reason for the
address liability issues of AVs. Other laws related to accident is a system failure or an inability of the AV to
computers are only applicable to commercial transactions react in a certain situation (Hevelke and Nida-Rumelin
and thus cannot be applied to AVs. Their solution is to 2015). In conventional vehicles in which any vehicle
consider AVs as people’s best friends. Tort law treats dogs malfunction is considered an error (e.g. a burst tire or
and computers as personal possessions as such the Laws system failure). While, an AV malfunction is a program-
that govern canine ownership demonstrates that the appli- ming error, which makes the above-mentioned parties
cation of strict liability to AV owners accomplish twofold involved in the manufacture of the AV liable (Geistfeld
purpose: identifying and establishing the liability of AVs 2017).
without hindering the mass adoption of the technology. But For shared AVs, Webb (2016) claimed that autonomous
there are no laws that clarify liability in accidents involving technology also impacts the insurance industry, and this
AVs. Since there is no law to guide courts regarding AV topic has generated much research. When the vehicle is in
liability issues, canine ownership laws can be applied to fully autonomous mode at the time of collision, insurance
AVs in an effective and practical way. Some clauses in companies shift liability from the driver to the manufac-
current vehicle and computer laws can also be applied to turer. One solution suggested is to establish a national fund
AVs. to pay for accidents involving AVs. This fund would be led
Marchant and Lindor (2012) affirmed that AVs, partic- and provided by manufacturers and would allow them to
ularly those that are privately owned, will change the way negotiate with other stakeholders regarding contributions,
liability is determined. Considering these dynamics, it is which would be a better option than going to court and
important to differentiate between partially and fully being sued.
autonomous cars. A partially autonomous car has systems The increase in ride-sharing services and the emergence
that warn drivers to reduce speed, for example, or to of AVs will change the nature of vehicle ownership and
change lanes before the vehicle touches another car or driver control, which underlines the need to change current
object (Gurney 2016). For partially autonomous cars, only traffic laws. Walpert (2016) raised some liability questions
some liability shifts from driver to vehicle, since the main for shared AVs. The first is what is the tort structure
decision-maker is still the human driver; nonetheless, col- appropriate for this imminent occurrence? The second
lision risks are reduced because of system interference and question is this: is current tort jurisprudence sufficient to
guidance. A fully autonomous car, however, takes full guide judges who will preside over accident litigation in
responsibility for any collision, since it takes charge of the the future? Walpert’s third question is how would legisla-
entire decision-making process, with no human interfer- tures faced with a growing number of AVs and increased
ence at all. ride-sharing modify current law? The answers to these
Furthermore, Marchant and Lindor (2012) stated that in questions will have serious implications for vehicle tech-
the case of a failure in the accident avoidance system, nology manufacturing stakeholders and for individual
partially and fully autonomous vehicles carry the same vehicle drivers. Apportioning liability could hinder AV
liability implications. Apportioning liability is difficult technological advances, even though AVs provide safety
when the human driver is given an opportunity to control benefits over conventional modes of transport.
the AV to avoid a collision but chooses not to use it. By Uber envisions having a fleet of AVs that transport
contrast, when the human occupant decides not to follow passengers without the need for human drivers, thus
the AV manual diligently (e.g. by using the AV in certain increasing operational efficiency (Casey 2016). The
weather even though one is clearly instructed not to do so), increased usage of vehicle-sharing services like Uber is a
liability should be apportioned to him or her, at least disruptive force in the automotive industry amid the

123
1230 Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249

emergence of AV technology (Kang et al. 2017). In vehi- 1.1 History of autonomous vehicles
cle-sharing services, drivers use their own private cars and
do not have commercial licenses (Shaheen et al. 2012). According to Pendleton et al. (2017), vehicle automation
This fact has raised issues under ride-sharing laws, such as was envisioned in 1918. General Motors produced the first
questions about the liability of drivers for tortious behav- concept, in 1939 (Shladover 2018). The technological
iors (Walpert 2016). developments that led to modern-day AVs started in 1948
On the issue of shared AVs, Walpert (2016) described when cruise control was invented (Ross 2014). The tech-
how individuals can be held liable for any outcome nology has continued to evolve over the last 70 years,
resulting from a shared experience. To do this, he applied leading to the eventual conception and introduction of
enterprise liability theory (Choi 2019), which proposes that AVs. Advanced information technology companies, such
separate entities can be liable for shared actions (Smith as Google started a driverless car project in 2009 and
2017). Liability can be ascertained by courts in cases when introduced such a car in 2010 (Urmson 2016).
plaintiffs are not able to show or prove any particular Automobile companies have been actively engaged in
entity’s fault (Abraham and Rabin 2019a). research and development to advance AV evolution. Volvo
Enterprise liability theory states that activities with tested its first AV in 2017 and plans to market it in 2021,
foreseeable hazards should bear the costs of the danger while BMW, Audi, Nissan and Mercedes-Benz are work-
they cause (Cowger 2018). Proponents expect enterprise ing to introduce their AVs in the market by 2020 (Faisal
liability law to move beyond traditional negligence and et al. 2019). Such clear plans from automobile companies
include no-fault negligence (Eastman 2016). Selbst (2019) require a massive transition in many areas, including in
argued that enterprise liability should make the doer bear laws and regulations.
the loss rather than distribute the cost based on the pro-
portions of fault. Since accidents cause injuries, enterprise 1.2 Appropriation of liability with the introduction
liability should focus on compensating the victim (Patti of autonomous vehicles
2019).
The Geneva Convention on Road Traffic (1949) and the Fast-paced developments have motivated researchers to
Vienna Convention on Road Traffic (1968) involve a large study topics associated with AVs, including AVs’ history,
number of countries and outline the effects of traffic laws evolution, classification and mass adoption; the challenges
on these countries. The Geneva Convention, signed by 96 in producing them; the attitudes and profiles of drivers
countries, was replaced by the Vienna Convention, signed influencing the adoption process; the implications of
by 73 countries. The countries that were signatories to the technological advances; and the laws and policies gov-
Geneva Convention but did not sign the Vienna Conven- erning their use.
tion are still bound to follow the former. These two con- The issue of civil and criminal liability has been dis-
ventions formulated and established traffic laws that should cussed extensively in the literature (Browning 2014). The
be followed by signatory countries to ensure traffic safety. driver is generally considered liable when a conventional
If the contents of these conventions are not updated in car is involved in an accident (Collingwood 2017). With
response to technology advancements, the conventions significant changes in conventional transport and infras-
may hinder the mass implementation of AVs, which are tructure imminent, current laws and regulations may not be
expected to reduce crashes due to human error, which is the applicable to the new situation. Various studies on this
cause of 90% of car crashes (European Parliament 2016). subject have argued that current laws need significant
For more than a decade, countries have been updating revision before the mass adoption of AVs.
UN regulations to incorporate technological advancements. The question of who is liable and responsible when AVs
The arrival of AVs will require further changes in these fail is a topic widely discussed in the literature (Cohen
regulations. The UN Economic Commission for Europe et al. 2017). The literature also identifies stakeholders,
(UNECE) has two bodies working to revise the UN con- including automobile companies, software manufacturers,
ventions: WP.1 (Global Forum for Road Traffic Safety) owners of AVs, insurance companies, occupants, policy-
and WP.29 (World Forum for the Harmonization of makers and lawmakers. However, for a crash involving a
Vehicle Regulations). Advancements, such as emergency driverless vehicle, it is unclear how liability would be
brakes and warning systems, form the basis for the updates apportioned among these parties and whether drivers would
in these conventions (UNECE 2012). Current treaties will liable if there were manual steps they could have taken to
require extensive amendments to account for AVs, since avoid the accident.
the existing conventions are based on traditional traffic in Taeihagh and Lim (2018) studied the steps in the law-
which humans are in full control (European Parliament making process in countries around the world and con-
2016). cluded that governments are currently amending laws and

123
Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249 1231

making new rules to govern the liability of AVs. The design and manufacture of AVs, including the vehicle
authors confirmed that in the US laws have been introduced manufacturer and the software company that produced a
to address privacy and cybersecurity, whereas the UK and malfunctioning AV component (Taeihagh and Lim 2018).
Germany have addressed liability issues related to AVs. There is currently no clear legislation that governs AV
These governments has taken two steps: amending current liability generally. Nonetheless, Fagnant and Kockelman
laws and creating new laws. These steps involve the cre- (2015) recommended that liability regulations should cre-
ation of working groups that address legal issues and make ate a balance among all parties involved without hindering
new regulations. technology adoption or the mass adoption of AVs. They
Geistfeld (2017) studied tort laws in the US and how also, added that the absence of a legal framework makes
these should be revised with the adoption of AVs. The the assignment of liability uncertain in the event of an
author recommended that the US government formulate accident.
two principal regulations on adequate premarket testing Bartolini et al. (2017) provided another recommenda-
and disclosure of safety risks by manufacturers, including tion. Since AVs are still in the testing stage, they argued, a
automobile and software companies (Aleksandrovskaya possible solution is to focus on the occupant’s role in
et al. 2019). A manufacturer should be apportioned liability avoiding an accident. They recommended that future laws
for a crash only if fault is found in the design or in the governing AVs should limit liability or impose no liability
company’s compliance with the regulations. Lohmann for human occupants even if those occupants did not act to
(2016) argued that with the adoption of AVs on the streets, avoid a crash. Possible interventions include adding an
humans should no longer be liable for any crash, and only alarm to warn the driver about traffic conditions so that he
the manufacturers should be blamed for any damages or she can take control of the vehicle. The driver can still
caused; this would represent a significant shift in current be free from liability since the idea behind an AV is that it
tort laws. is driverless, and the vehicle takes control in all situations.
Genziuk (2015) discussed the vehicle owner’s respon- Nonetheless, failure to warn may be a fault on the part of
sibilities in different scenarios in which the liability for a the AV, making manufacturers liable.
crash is difficult to apportion. Whether occupants should be Sun et al. (2016) recommended that liability be removed
blamed in cases where they could have taken steps to avoid completely from the driver and shifted to the vehicle
the accident is another topic to consider. The human ability manufacturers. This can make certain decisions challeng-
to make prompt decisions cannot be ignored even with the ing, particularly as to whose safety should be prioritized:
advent of modern technology, and this aspect must be the AV occupants or others using the road (e.g. pedestri-
considered when revising current tort laws across the ans). Moreover, since drivers can override the AV controls
world. Moreover, most governments make manufacturers (e.g. in a heavy traffic situation), the issue of shared lia-
liable when there is a critical fault found in an AV, as these bility arises. Additionally, for fully automated vehicles,
require repair and maintenance time (James et al. 2018); there are situations when the car is made to follow the
however, the liability of the driver remains undetermined. passengers’ instructions (e.g. picking the passenger up
The way liability is assigned affects insurance costs from a location despite the presence of weather conditions
(Abdullah 2016) and changes the overall dynamics of under which it is recommended the AV not be driven). Sun
automobile insurance. While the insurance industry con- et al. (2016) concluded that there are currently no legal
templates the impact of the adoption of AVs, revised laws frameworks regarding AVs and that huge changes are
are needed to cope with possible increased costs, as the expected in the fundamental components of liability in new
liability for accidents will be assigned differently. legislation.
The key problem faced by AVs is that the laws used by AVs should be tested for millions or billions of kilo-
most countries address only conventional traffic, in which meters to confirm their reliability and safety. Even after
the car is considered an object under human control, thus aggressive testing, AVs would need hundreds of years of
putting the liability on the driver only. Since driving is driving over many kilometers to ensure reliability. Satis-
considered a dangerous activity that puts others’ life at risk, factory testing of AVs is not currently possible, and inno-
there are strict laws governing conventional traffic. The vative methods to gauge safety must be developed
legal framework for conventional vehicles includes man- (Baglietto et al. 2018). The reliability of AVs has been
ufacturing requirements, administrative requirements for questioned by Kalra and Paddock (2016) through the fol-
car ownership (e.g. license and car registration), insurance lowing three questions:
requirements, liability principles, and driving rules and
1. For how many kilometers should AVs be tested
regulations (Bartolini et al. 2017).
without fail to compare their failure rate against given
There is no definite regulatory framework that defines
benchmarks?
how liability is placed on third parties involved in the

123
1232 Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249

2. For how many kilometers should AVs be tested can help determine who is at fault. Despite issues of pri-
without fail to compare their failure rate against given vacy, the black box is expected to be used to identify the
precision degrees? causes of crashes (Dhar 2016).
3. For how many kilometers should AVs be tested In accidents involving conventional vehicles, the driver
without fail to compare their failure rate against that is in full control of the vehicle and is thus held liable; in
of human drivers? AVs, by contrast, humans are not in control and cannot be
held liable (Douma and Palodichuk 2012; Collingwood
According to Kalra and Paddock (2016), the answers to
2017). As such, all or some of the liability will be shifted to
these questions require statistical approaches. The
AVs. Accordingly, road accidents can present issues
researchers provided formulas to answer these questions
relating to product safety such that third parties involved in
that can be used as references to test the reliability of AVs
the design and manufacture of AVs will be held liable and
statistically, as shown in the Table 1 below.
will face lawsuits (Marchant and Lindor 2012). The divi-
Table 1 presents the results of Kalra and Paddock’s
sion of liability between humans and technology is still
(2016) analysis. The table lists the three questions and the
unclear. The questions of whether humans should be held
corresponding results (i.e. fatality rates, reported injury
responsible for an accident when there is a chance of taking
rates and reported crash rates). The results show the
control of the vehicle to avoid an accident also remains
number of kilometers and number of years (or months)
(Collingwood 2017). Additionally, there is no clear regu-
needed to test AVs to reach the given failure rates in the
latory framework on how liability should be apportioned
first column. It is clear that AVs must be driven for billions
among third parties involved in the manufacture of AVs
of kilometers to reach these benchmarks, distances which
(Collingwood 2017).
are impossible to achieve in public road tests. Additionally,
Given the above points, there will be higher risks for
improvements in AVs will also need to be tested for mil-
manufacturers of AVs involved in a crash, and these can
lions of years and kilometers to ensure performance and
affect their brand and reputation (Hevelke and Nida-
reliability. These results have implications for liability and
Rumelin 2015; Tien 2017). The current regulatory frame-
regulatory laws, policymakers and the public (Baglietto
work does not identify the moral responsibility of pro-
et al. 2018).
grammers in their designs of life-and-death decisions
(Fleetwood 2017). Policymakers have yet to standardize
1.3 Liability issues and regulatory framework
decision-making criteria for algorithms—for example,
for autonomous vehicles in selected countries
whether decisions should be made based on priority,
severity, type of injury or number of injured victims
In their research, McChristian and Corbett (2016)
(Fleetwood 2017). Only the UK has revised its regulatory
explained that the driving in an semi-autonomous vehicles
framework to include these issues related to the liability of
is shared between the vehicle and the human occupant, and
drivers and third parties (Duffy and Hopkins 2013; HC 143
this complicates liability. The transition from conventional
2017).
traffic to automated driving will mean insurance companies
There is currently very limited knowledge on the effects
must depend on ‘‘black boxes’’ to monitor the driving.
of shifting liability on insurance costs (Baumann et al.
Similar to those used in airplanes, the black box in an AV
2019). In case of an injury, parties can sue the AV

Table 1 Kilometers and years needed to establish the reliability of autonomous vehicles (AVs)
For how many kilometers or years should AVs Benchmarks for failure rate
be tested?
A. Fatality rate: 1.09 B. Reported injury rate: 77 C. Reported crash rate: 190
fatalities per 160 million reported injuries per 160 reported crashes per 160
kilometers million kilometers million kilometers

To have 95% confidence in their failure rate 275 million kilometers 3.9 million kilometers 1.6 million kilometers
(12.5 years) (2 months) (1 month)
To have 95% confidence in their failure rate 8.8 billion kilometers 125 million kilometers 51 million kilometers
being within 20% of the true rate (400 years) (5.7 million years) (2.3 years)
To have 95% confidence and 80% power that 11 billion kilometers 161 million kilometers 65 million kilometers
their failure rate is 20% better than the failure (500 years) (7.3 years) (3 years)
rate of humans
The time is assessed with a fleet of 100 AVs driving 24 h a day at a speed of 40 kph

123
Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249 1233

manufacturer, especially if there is system fault. This identify the eligibility of crash victims (NTC 2017). The
increased liability can affect the motivation of manufac- governments of both Singapore and Australia have made
turers to innovate, which can slow the rate of safety gradual but consistent efforts toward formulating a set of
improvement of AVs (Hevelke and Nida-Rumelin 2015). regulations based on light control-oriented strategy.
The US government has delegated the responsibility for The Chinese and South Korean governments have yet to
identifying liability to state governments (NHTSA 2017). identify their regulatory stance toward liability and insur-
The US Department of Transportation has not shown any ance, which puts them in the no-response strategy category.
interest in formulating laws for liability and insurance. The Dai (2018) urged the Chinese government to formulate
National Highway Traffic Safety Administration (NHTSA) laws for AV testing. The South Korean government rec-
encourages states to allocate liability and formulate insur- ognizes that the absence of international standards is a
ance policies. Most states in the US have started addressing barrier to formulating local laws on AVs, and since the
the liability aspects of AVs (NHTSA 2017). country imports and exports cars, it is imperative that
The UK is the only country that has implemented a manufactures comply with international standards
toleration-oriented strategy to handle liability and insur- (Ramirez 2017).
ance risks, in which policy-makers act to confirm that the The EU has still not revised its laws on AV liability and
system or organization’s performance is robust to risks insurance risks and is still in the process of exploring
under a wide range of conditions. Whereas no-response is options. The European Commission launched GEAR 2030
when policy-makers do not take steps to address risks and in 2016 to evaluate options for AV-related laws and made
may delay decisions due to their uncertain nature or con- recommendations in 2017. European Parliament members
trol-oriented strategies. A control-oriented is defined as suggested that the EC should introduce an insurance
policy-makers accept for the existence of threats, however scheme, along with a fund, to ensure compensation for AV
take steps to control them by executing official policies and crash victims and identify the liability for all the parties
regulations and are implemented by other countries involved (EPCLA 2016; European Parliament 2017).
(Taeihagh and Lim 2018). The Center for Connected and Taeihagh and Lim (2018) stated that, similar to the UK,
Autonomous Vehicles identified the gaps in liability- and Germany formulated permanent legislation in 2017 to
insurance-related laws and proposed amendments to the address the liability and insurance aspects of AVs.
UK Department of Transport in 2016 (CCAV 2016). A bill According to these laws, AVs are required to install a black
(HC 143 2017) was passed in response which clarified the box that is programmed to record the car’s journey, and it
liability of insurers and AV owners in cases of accidents should be used as reference in cases of crashes (JDSUPRA
and other circumstances. Insurers are made liable for death 2017; Wacket et al. 2017). The laws ensure a balanced
or damages caused by insured AVs (HC 143 2017). apportioning of liability between the manufacturer and the
Insurers are, however, only partly liable in cases when the driver, with the former taking liability when AV technol-
AV owners are found at fault. The bill thus has ambiguities ogy is at fault and when there is a system failure (Wacket
regarding the appropriation of liability to the parties et al. 2017). However, the laws do not clarify the maximum
involved. The bill follows existing motor insurance amount of time drivers are given to take control in case of
schemes but ensures faster compensation for injured parties an impending crash nor do they explain issues related to the
(CCAV 2016). The Consumer Protection Act protects ownership of black box data (JDSUPRA 2017). For this
manufacturers if they are able to prove that an AV was not reason, Germany’s Ethics Commission formulated and
defective when it was delivered to the owner and that the published guidelines for AVs, which recommend that it be
fault was identified later because of technological made clear who is considered the driver and who should be
advancement. authenticated for determining liability. The commission
Singapore and Australia recognize the need to revise also considered it unethical for algorithms to use individ-
liability laws governing AVs. The Singaporean govern- uals’ data (e.g. age and gender) when making decisions to
ment revised the Road Traffic Amendment (RTA) in 2017, avoid an accident (Di Fabio et al. 2017). Although not
which exempted AVs and parties involved in AV testing mandatory, these guidelines can be considered a first step
from existing clauses of the RTA, which hold humans toward an ethical framework for issues related to AVs.
liable (CNA 2017). Singapore clearly acknowledged that Further discussions are needed to determine the responsi-
human control is diminished in AVs and that the use of bility of persons designing algorithms.
AVs requires changes in the existing road and criminal Japan’s strategy in formulating laws regarding liability
laws (MOT 2017). The Australian government set deadli- and risk is categorized light control-oriented (Taeihagh and
nes to revise its liability and insurance laws by November Lim 2018). The National Police Agency has recommended
2017. The government also committed to evaluate current actions to reduce risks, but these are not mandatory. For
laws, identify liability aspects and revise injury systems to instance, manufacturers are recommended to install black

123
1234 Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249

boxes in AVs that are being tested (Nikkei 2018). More- mandated member countries and industry players identify
over, AV testers should submit documents regarding the joint goals to encourage the operation of AVs (European
AV’s structure and risks to help authorities make plans for Union 2016). The future regulatory framework is generic in
responding to accidents. Meanwhile, AV operators should nature, stating that laws should support the use of AVs by
have a driver’s license and should take over responsibility ensuring flexibility. It identifies steps to facilitate AV
for the vehicle in the event of operational failure (Jiji 2017; introduction and implementation. The EC believes that
Japan Bullet 2017). Liability is placed on manufacturers in existing laws and regulations are sufficient to enable the
cases of errors and system defects, but programmers or any AV evolution in transportation (European Union 2009).
other third party involved in the AV design are excluded In his research, West (2016) explained that each country
from liability (Japan Bullet 2017). must address specific issues (e.g. policy, budget and ethics)
McChristian and Corbett (2016) stated that policymak- to enable the progress and development of AVs. In China,
ers around the world confront issues related to liability. for instance, there is a need to formulate a national policy
The UK’s Department of Transport reviewed existing laws on AVs. China has multiple ministries responsible for the
for AVs and found that these do not hinder AV testing on regulation and supervision of AV driving, and greater
public roads. Similar reviews have been done in Germany clarity is needed regarding who is responsible for what
and Sweden. aspect to establish a definitive framework. The government
According to Alessandrini et al. (2015), many countries also needs to invest more in infrastructure, relax regula-
around the world want to be a leader in the AV industry. tions on AV testing and limit restrictions on AV
Google has received permission to conduct testing for its manufacturers.
driverless cars in many US states. Similarly, Volvo has Meanwhile, in Europe, companies work to strengthen
received permission to test its AVs in Sweden. AVs are the artificial intelligence capabilities of AVs. Firms such as
also tested in Germany and France—both countries have a Google have entered the transportation industry to diver-
10-year roadmap to becoming a leader in AVs. In the UK, sify and develop programming skills in the sector. Auto-
huge investments in infrastructure development were made mobile manufacturers such as BMW and Audi compete in
in 2013 to support AVs, with the aim of making the country a fiercely competitive market that requires their AVs to
a leader in the driverless car industry. The UK also plans to have strong artificial intelligence and computing features,
review its laws to make the country a place where manu- since vehicle performance is based not only on the design
facturers can implement, develop and test AVs. Honda, but also on the software. The European Union also
Mercedes, Audi, BMW and other manufacturers plan pilot encourages data protection and privacy regulations,
projects, whereas Google is already in the testing phase. because information about people’s movement and loca-
Both IBM and Cisco aim to be suppliers of AV software. tion are made available to third parties.
Tire producers, such as Michelin, have been supportive of In Japan and Korea, governments and manufacturers are
the development of AVs. Renault and Tesla have also been more cautious (West 2016). Firms such as Toyota and
part of AV development. Honda invest hugely in the development of AVs even
Stocker and Shaheen (2017) stated that in the US, both though they are still unsure whether AVs should be a pri-
the federal government and state governments are in the ority and whether they should invest in new technology to
process of formulating laws governing AVs. The NHTSA be a part of this advancement. If they fail to keep up with
has released the Federal AV Policy, which forms the basis advancements in AVs, they will be left behind when AVs
of the regulatory framework for AVs in the US, proposing are mass produced in the future. In the US, the major
checklists on safety, privacy, ethics, cybersecurity and challenge lies in the integration of efforts among the 50
other issues. The document serves as a guideline for federal states, since they all have their own preferences on
and state governments when formulating laws. The policy licensing and regulations. Car manufacturers, such as Ford
suggests that state governments should split the responsi- and General Motors, encounter confusing rules in different
bilities that should be managed by federal and state gov- states, which hinders their progress. Firms are unable to
ernments. The NHSTA suggests that states should register innovate because of unclear and conflicting laws in the US,
AVs and regulate roads and safety measures, while the and greater clarity is required on liability and other issues
federal government should regulate AV performance. related to AVs.
Some states, including California, Michigan and Wash-
ington, have formulated their own AV guidelines, while 1.4 Liability issues with autonomous systems
others have passed executive orders (National Conference in other sectors
of State Legislatures 2016).
In Europe, the European Commission proposed the According to Barfield (2018), robots have become
Declaration of Amsterdam in 2016 (Patti 2019), which increasingly independent from human control, particularly

123
Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249 1235

amid advancements in artificial intelligence that allow to determine whether liability should be assigned to the
robots to do algorithm-based problem-solving. Autono- manufacturer or the driver (Schellekens 2015).
mous robots are capable of undertaking high-level goals Moreover, Perc et al. (2019) argued that since tort law
and can identify ways to achieve those goals with little or aims to reduce the number of accidents, policymakers
no human control. Nowadays current laws do not hold should adopt tort liability standards in determining liability
robots liable for any damage they cause with existing for accidents caused by intelligent objects, as this would
technology, regardless of how autonomous the robots are. encourage more investments in research and development
When determining liability for any accident caused by a to improve human safety. Most injuries are caused by
robot sold after completing tests and trials, courts try to negligence—a situation in which a party is liable for its
determine whether there was any defect in the robot that own unreasonable conduct (Ahmed 2019). If no miscon-
caused the accident. Product liability is determined under duct has been identified and the standard has been fol-
tort law, which identifies the liability for any damage lowed, the incident can be declared as ‘‘a matter of
caused by products sold to customers commercially, such chance’’, and no one can be made liable for it. In cases
as vacuum cleaners (Abraham 2017). involving computers, a strict liability standard is applied,
On the other hand, under tort law, liability is placed on which means that liability is assigned to the manufacturer
manufacturers of robots involved in any accident causing regardless of whether someone is at fault. Such strictness
damage even though robots are not directly causing it. This causes barriers for automation manufacturers, since
clause is known as strict liability, and this tort law aims to machine-controlled products causes greater liability for
regulate manufacturers’ activities that can cause risks to manufacturers Additionally, existing laws negatively
society through their products (Abraham 2017). Product impact technology advancements aimed at improving
liability refers to manufacturers being held liable for pro- safety (Ebnali et al. 2019).
ducing and distributing a dangerous or defective product
that can cause damage when used (Barfield 2018). It is yet 1.5 Liability of different AV automation levels
to be determined whether an autonomous robot can be and consumer adoption timeline
considered a dangerous product, since such a robot incor-
porates significant technological advancements. When In 2017, European, American and Australian authorities
customer injury results from a defective product, strict formulated a common language called the Society of
liability is placed on manufacturers, and they must pay Automotive Engineers (SAE) International Standard J3016
compensation to victims. Laws require manufacturers to (SAE International 2018) for identifying AV capabilities.
guarantee that their products are safe for public consump- The SAE identified six levels of automation from Level 0
tion and that they do not pose risks to consumers. These to Level 5, with Level 0 indicating no automation at all and
clauses of the tort law put all the liability on manufacturers, Level 5 signifying full automation. The categorization
so they engage in extensive research and development and depends on the way in which steering and braking are
make sure that their products undergo tests and trials before controlled—the extent to which human control is needed
market launch. The law holds manufacturers liable without when driving and whether the AV can operate without
considering the amount of effort the manufacturer devoted human control in all possible situations. The six levels
to developing and testing a product and regardless of identified by the SAE are as follows:
whether the customer had a contract with the manufacturer
– Level 0—No Automation: The human driver controls
(Abraham 2017).
all the driving.
Consilvio et al. (2019a; b) stated that faults related to
– Level 1—Driver Assistance: The system provides
circuits can occur suddenly, and there is no way to monitor
assistance in one task, such as steering or braking,
the condition of a circuit; nevertheless, board and relay
while the human driver performs all the remaining
faults are included under tort liability. Perc et al. (2019)
tasks.
affirmed that although manufacturers engage in extensive
– Level 2—Partial Automation: The system provides
attempts to perfect their products and ensure 100% relia-
assistance in both steering and speed control, while the
bility to avoid any liability issue (Kirkpatrick 2013), failure
human driver is required to monitor the environment
in robot technology may occur and cause damage or injury.
and act accordingly.
In such cases, tort liability is applied, and manufacturers
– Level 3—Conditional Automation: This is Level 2
are held liable. Some difficulties in tort liability are
automation with the additional feature that the human
expected in the case of AVs: for example, a situation in
driver is not required to monitor the environment when
which a driver thinks that the AV is engaged in an unsafe
the car is in automated mode. The driver, however,
action and takes control of the vehicle when in fact the AV
needs to respond to requests from the AV.
is operating in a safe manner. In this situation, it is difficult

123
1236 Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249

– Level 4—Highly Automated: This is Level 3 with the It is expected to take several decades before AVs totally
additional feature that no human intervention is needed replace conventional vehicles and become the only type of
when the car is in automated mode. vehicles allowed on public roads. During the transition,
– Level 5—Fully Automated: The car is fully automated accidents involving conventional cars will continue, in
and in control in all conditions and environments. addition to accidents involving motorcycles, pedestrians
and/or third parties. According to Abraham and Rabin
AV manufacturers decide which options to provide for
(2019b), some accidents involve conventional vehicles are
human drivers so they can take back control from the
caused by defects in the automated features. But even
system in Levels 4 and 5 of automation. Table 2 summa-
though these vehicles have certain automated features,
rizes the six levels.
current regulations on driver liability and product liability
The extent of automation, the requirements for the
will remain unchanged.
human driver in terms of controlling the vehicle and
Shavell (2019) opined that since the level of automation
monitoring the driving environment, and the ability to
is increasing, drivers will become less liable for crashes.
control the automated system can all impact society and the
The major tort rule is determining the fault of the driver,
regulations it imposes. The degree of automation defines
and this will become irrelevant when drivers are no longer
the tort duties of individuals operating the AVs, since the
in control of the vehicles. The faults of AVs will shift the
SAE categorization considers the extent of human inter-
liability away from the human driver; hence, policymakers
vention in the AV operations (Walpert 2016).
and the society will focus more on the liability of vehicle
Future tort law will develop from the current system of
manufacturers. Manufacturers will be required to pay
apportioning liability for car crashes (Abraham and Rabin
compensation for damages caused by AVs, especially
2019b). When Level 2 and Level 3 vehicles are involved in
damages caused by defective components. Nevertheless,
accidents, they are considered under accident laws for
liability can also be placed on vehicle owners for their
conventional vehicles, since significant human control is
inattention to these defects or for their failure to upgrade a
involved. As these levels become more automated due to
system.
advancements in technology, tort laws must be revised and
Current tort liability laws will continue in effect for
amended to address issues involving Level 4 and Level 5
some time, since conventional vehicles are still the major
vehicles. In Levels 2 and 3, driver control depends on the
type of vehicle on public roads. As a result, the liability of
need for safe operations; drivers therefore need to be aware
human drivers will also remain as part of the current
of the environmental situation so that they can take control
practice of considering human error in determining
when necessary (Merat et al. 2014).

Table 2 Different levels of vehicle autonomy according to the SAE standard


SAE Description Execution of Monitoring of Fallback System
level steering and driving performance of capability
acceleration/ environment dynamic driving (driving
Deceleration task modes)

Human driver monitors the driving environment


0 Only the human driver controls all the driving Human driver Human driver Human driver n/a
1 The system provides assistance in one task (e.g. steering Human driver and Human driver Human driver Some driving
or braking), while all remaining tasks are performed by system modes
the human driver
2 The system provides assistance in both steering and speed System Human driver Human driver Some driving
control, while the human is required to monitor the modes
environment and act accordingly
Automated driving system monitors the driving environment
3 Level 2 automation with the addition that the human System System Human driver Some driving
driver is not required to monitor the environment when modes
the car is in automated mode. The driver, however,
needs to respond to AV requests
4 Level 3 with the addition that no human intervention is System System System Some driving
needed when the car is in automated mode modes
5 The car is fully automated and in control in all conditions System System System All driving
and environments modes

123
Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249 1237

responsibility for an accident. Nevertheless, technological 2018). However, further investigation revealed that Uber
advancements in conventional vehicles can significantly disabled the emergency braking system to avoid erratic car
reduce accident rates (Woods 2019). Existing liability behavior (National Transportation Safety Board 2018b).
regulations focus on drivers, since human errors cause the Uber switched off the safety feature when the driver was
majority of accidents. Negligence liability is applied to car not in control; the feature was kept on when the driver was
crashes involving conventional vehicles. Third-party in control (Laris 2018). This disabled emergency braking
insurance protects drivers, whereas underwriting focuses function caused the crash. If it had been kept on, the sen-
on both owners and drivers (Chatzipanagiotis and Leloudas sors could have detected the collision object, and the crash
2020). could have been avoided (Hawkins 2018). Moreover, the
The main concept in Level 5 automation is noninter- car should have applied the brakes when it was 20 m from
vention and noninvolvement of the human driver, which the individual (Krisher 2018). The driver was not impaired,
means that AVs cannot count on the human driver to and weather conditions were also ruled out as the cause of
provide inputs related to vehicle operation (Koopman and the crash (Daisuke 2018).
Wagner 2016). Boeglin (2015) opined that the parties who In another incident in 2018, a Tesla Model X crashed
should be held liable for a crash involving AVs are as into a barrier in California and killed the driver. The
follows: the AV owner, the AV user, the AV manufacturer, investigation found that the driver was kept alert during the
the manufacturers of AV components and government drive with visual and auditory alert systems, but he was not
bodies (Ryan 2017). Since these parties act interdepen- alerted for 15 min before the accident occurred. The driver
dently to ensure the effective operation of AVs, it will be also did not handle the steering for the 6 min leading up to
difficult to ascertain individual levels of responsibility. the accident (Bomey 2018). Moreover, the speed increased
Boeglin (2015) added that privacy tools like the black box from 100 to 112 kph just 3 s before the crash with no
can shift the liability from owners and users to third parties braking or steering movement detected (National Trans-
(e.g. manufacturers or government bodies). portation Safety Board 2018c). The crash was considered
A reasonable timeframe for the adoption of AVs partially the fault of the driver and partially the fault of
remains unclear. However, Stone et al. (2020) affirmed that Tesla. Tesla was considered responsible for the driver’s
with the mass adoption of AVs, questions related to con- inattentiveness, since he relied solely on the autopilot
sumer adoption and policy changes must be addressed. If system (National Transportation Safety Board 2017). In
technological advancement continues at the current pace, this example, the system caused the driver’s inattention,
15 to 20 years will be needed before higher-level AVs increasing his risk for crashes. Drivers are thus advised to
appear on public roads (Beiker 2016). The Netherlands remain attentive even when the vehicle is on autopilot.
Institute for Transport Policy Analysis (KiM 2015) pre- This paper presents a systematic review of the literature
dicted that a complete transition to high-level automation on liability with AVs. Section 2 explains the methodology
will occur approximately between 2060 and 2100 and that used for the systematic literature review, and Sect. 3 dis-
Level 5 automation will not be achieved in the near future. cusses the results and findings of the review. Section 4, the
last section, presents the conclusions.
1.6 Fatalities involving autonomous vehicles

According to Ackerman (2019), 2018 saw an increase in 2 Methodology and data collection procedure
the number of car crashes involving AVs causing fatalities.
For example, an Uber self-driving car in Arizona killed a The research presents a systematic review of existing lit-
pedestrian (Daisuke 2018), and a Tesla on autopilot cra- erature on liability with AVs. A systematic literature
shed into a barrier, killing its driver (National Trans- review involves the search, review and analysis of relevant
portation Safety Board 2018a). Soo (2018) also reported articles on a topic to reach conclusions. Bacca et al. (2014)
another Tesla on autopilot crashing into a fire truck in identified the three main phases of a systematic review:
Utah; the driver survived the incident. Meanwhile, the planning, conducting the review and reporting the findings.
Californian government received 129 AV collision reports, This method was used in the present paper to systemati-
which included AVs hitting barriers and pedestrians and cally derive and analyze the data to formulate conclusions.
AVs stopping in traffic; these reports increased safety Figure 1 presents the methodology used in this paper.
concerns (State of California 2020) and instilled greater Planning involved clearly formulating the research
awareness regarding AV-related issues. questions and determining how they can be answered
In Arizona, a Volvo Uber self-driving car killed a person through a systematic review. The research question of this
(Kerr 2018). The video footage showed that the driver was paper aims to answer is as follows: How has past literature
on her phone and was not handling the steering (Daisuke addressed the topic of liability with AVs? The objective is

123
1238 Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249

The inclusion and exclusion criteria were established


Formulating research questions and used to sort the articles, identify common character-
istics and choose the most relevant data. An article that is
25 years old cannot be compared with an article published
a year ago, as this can create discrepancies in the results.
Well-defined inclusion and exclusion criteria can help sort
Identifying keywords and databases the data on a uniform scale, which makes the analysis
easier since the background and context of the articles are
the same. Table 3 presents the inclusion and exclusion
criteria for the present research.
The inclusion and exclusion criteria enabled the sorting
Identifying inclusion and exclusion of articles according to relevance, ensuring that the final
criteria articles selected provided dependable findings. A total of
150 articles were identified using the keyword search of the
EBSCO, SCOPUS and PROQUEST databases. Only
English-language articles were chosen, since it is the lan-
guage easily understood by the researcher, and it is widely
Analyzing the articles
used around the world.
To select the most credible research, only peer-reviewed
articles were chosen for review. An article that has been
peer reviewed is more credible, since it has already been
Reporting results and reaching evaluated by one or more competent individuals who have
conclusions given their opinions. Of the 150 articles, 134 were peer
reviewed; the rest were thus excluded from the review. All
articles in EBSCO and SCOPUS are peer reviewed. In
Fig. 1 Methodology used in the present research
contrast, PROQUEST has some non-peer-reviewed arti-
to analyze previous studies and their conclusions regarding cles. Of the 71 articles identified, 16 were excluded
liability issues related to AVs. because they were not peer reviewed also, there was no any
The keyword method was used to search the databases time constraints for the articles collected. But the articles
for relevant articles. Since ‘‘autonomous’’ is synonymous collected was till the third quarter of the year 2019. Table 4
with ‘‘driverless’’ and ‘‘self-driving’’ when describing AVs, summarizes the process of identifying the included articles.
these terms were identified as keywords. Notably, the term Out of 134 articles, duplicates were removed, resulting
‘‘liability’’ was used with each of these three terms. The in 74 articles. Of these 74 articles, three were written in a
following search terms were thus used to search the data- language other than English (i.e. Polish and Portuguese
bases for relevant article titles, abstracts and keywords: languages); these were excluded because the researcher can
only read and understand the English language. The final
– Liability autonomous vehicles, article count was thus 71. The article list is presented in
– Liability driverless vehicles, and Appendix A. The relevant data from the articles were
– Liability self-driving vehicles. recorded in nine Excel spreadsheets, using three search
When conducting a systematic literature review, Bramer terms in three databases. The Excel sheets present a sum-
et al. (2017) advise using multiple databases. The present mary of the search by tabulating the article titles, author
study used three databases—SCOPUS, EBSCO and PRO- names, journal titles, publication years, volume numbers,
QUEST—to gather articles using the above keywords. The issue numbers, publishers, subjects, keywords, abstracts
articles were then sorted based on the inclusion and and page links. The results of the analyses are presented in
exclusion criteria. Lam and Kennedy (2005) state that the charts, graphs and percentages. The final 71 articles were
success of an analysis depends on the quality of data analyzed to explore the themes, subject areas, publication
gathered; hence, it is critical to identify the proper inclu- years, journal names and keywords found in their titles and
sion and exclusion criteria. According to Mallet et al. abstracts.
(2012), the inclusion and exclusion criteria for data
screening are used to achieve objectivity and to screen the
data in a uniform way.

123
Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249 1239

Table 3 Inclusion and exclusion criteria used in the research


No. Inclusion criteria Exclusion criteria

1 Peer reviewed Not peer reviewed


2 Non-duplicate articles Duplicate articles
3 English language Non-English language
4 Having one or more search terms in the title, abstract or keywords Not having any of the search terms

Table 4 Synthesis of the identification and inclusion of articles


Database Search terms Identified After deleting After deleting After deleting
Papers non-peer-reviewed duplicate papers non-english
Papers papers

EBSCO Liability autonomous vehicles 23 23 9 9


Liability self-driving vehicles 2 2 2 2
Liability driverless vehicles 2 2 2 2
SCOPUS Liability autonomous vehicles 35 35 21 20
Liability self-driving vehicles 12 12 8 8
Liability driverless vehicles 5 5 3 3
PROQUEST Liability autonomous vehicles 53 42 21 19
Liability self-driving vehicles 10 8 4 4
Liability driverless vehicles 8 5 4 4
TOTAL 150 134 74 71

3 Findings Also, it shows from the number of the articles published


that researchers, governments and policymakers started to
3.1 Publications over time consider and place attention to laws and regulations
regarding liability. Moreover, liability needs to be appor-
The analysis of data shown in Fig. 2 reveals that 45% of tioned before the mass adoption of AVs; nonetheless, the
the articles were published in 2018 and 2019, while 21% literature on AVs in recent years seems to understand the
were published in 2017. It is clear from this analysis that importance of the issue.
the quantity of literature on liability with AVs started
growing at a faster pace in 2015, after the successful 3.2 Journal subject areas
introduction and testing of fully autonomous vehicles and
the announcements by automobile companies that they Regarding the subject area, 35.2% of the articles (i.e. 25
would introduce their own AVs into the market by 2020. out of 71 articles) were found to be in the social sciences,
while 19.7% (i.e. 14 articles) were under computer sciences
18
and engineering. Business management and decision sci-
16 16 ence articles comprised 5.6% (i.e. 4 articles) of the total.
16 15
Based on this analysis, it can be said that research on lia-
14
bility with AVs has been conducted under a wide range of
12
No. of Arcles

subject areas and that researchers from multiple fields take


10 9
an interest in the topic (see Fig. 3).
8 7
6
3.3 Article title keywords
4
2 2
2 1 1 1 1
Analyzing the titles of the 71 articles (see Table 5) revealed
0 that 30 articles mentioned ‘‘autonomous vehicles’’ in their
2000 2006 2011 2012 2013 2014 2015 2016 2017 2018 2019
title, whereas 23 had ‘‘liability’’ or ‘‘liable’’ in their title.
Year
Moreover, 15 articles had one or more law-related key-
Fig. 2 Number of articles per year words, in their title such as ‘‘law’’, ‘‘legal’’, ‘‘policy’’ or

123
1240 Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249

Fig. 3 Distribution of articles


by subject area

Table 5 Keywords in the titles of articles reviewed published articles. Figure 4 also includes six journals with
two published articles. The other 47 journals published
Article title keywords No. of articles
only one article each; these journals are not listed in Fig. 4.
Autonomous vehicles 30
Liability or liable 23 3.5 Citation analysis
Law, legal, policy or regulations 15
Self-driving 7 Citation analysis is a method used to judge the performance
Ethics, ethical or ethico 6 of a previous research publication by determining the
Risk 5 number of later studies that cited the said research. When
Accidents 3 research is published, other researchers can refer to it in
their work. Citations in other studies indicate that the
previous research is reliable and has contributed positively
to the field. The 10 articles with the greatest numbers of
‘‘regulations’’. Seven articles had the keyword ‘‘self-driv- citations were selected for this analysis. Table 6 summa-
ing’’ in the title; six articles had ‘‘ethics,’’ ‘‘ethical’’ or rizes the citation analysis.
‘‘ethico’’; five mentioned ‘‘risk’’; and three mentioned Articles from 2013 and later were the most cited in later
‘‘accidents’’. research, indicating that current studies give more attention
Notably, researchers mostly discuss the legal aspect of to research conducted after 2013. Conversely, studies
the liability issue to help policymakers formulate suit- conducted before 2013 were not cited extensively in later
able laws and regulations to enable the smooth mass articles.
adoption of AVs and reduce any negative impact.
Researchers also discussed how liability would shift from 3.6 Papers content analysis
the drivers to other parties in the case of AVs. The Law-
related terms were found to be present in the titles of all A thorough analysis of the articles, including content,
articles that discussed the liability aspects of AVs. methodology, approach and findings, shows that most of
them are qualitative in nature, and their content mainly
3.4 Journal count analysis discusses different types of liability like tort liability,
product liability and strict liability in addition to, sustain-
As shown in Fig. 4, the journal count analysis indicates able transportation, cost-effective, data privacy, licensing
that the journal with the greatest number of articles configurations for AVs, travel time reduction, crash sav-
reviewed was published in Transport Research Part A: ings, parking benefits, fuel efficiency accountability and
Policy and Practice (i.e. five published articles). The the laws that need to be updated to encourage the public to
European Journal of Risk Regulation had four published adopt AVs. In addition, designing tort liability for vehicle
articles, whereas Transport Research Part C had three manufacturers in a way that will support the development

123
Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249 1241

Fig. 4 Journal count


Transportaon Research Part A

European Journal of Risk Regulaon : EJRR

Transportaon Research Part C

Northwestern Uni Law Review

Berkeley Technology Law Journal

Computer Law and Security

Science and Engineer Ethics

Transport Reviews

Journal of Muldisciplinary Research

0 1 2 3 4 5 6
No. of Arcles

Table 6 Most cited articles


No. of Article title Author(s) Year Journal
citations

1,322 Preparing a Nation for Autonomous Vehicles: Fagnant, D.J. and Kockelman, K 2015 Transportation Research
Opportunities, Barriers and Policy Recommendations Part A: Policy and
Practice
143 Responsibility for Crashes of Autonomous Vehicles: An Hevelke, A. and Nida-Rümelin, J 2015 Science and Engineering
Ethical Analysis Ethics
69 A Roadmap for Autonomous Vehicles: State Tort Liability, Geistfeld, M. A 2017 California Law Review
Automobile Insurance, and Federal Safety Regulation
58 Self-driving Cars and the Chilling Effect of Liability Law Schellekens, M 2015 Computer Law & Security
Review
51 Governing Autonomous Vehicles: Emerging Responses for Taeihagh, A., Lim, H. and Si, M 2018 Transport Reviews
Safety, Liability, Privacy, Cybersecurity, and Industry
Risks
47 Autonomous Vehicle Regulation: How an Uncertain Legal Brodsky, J. S 2016 Berkeley Technology Law
Landscape May Hit the Brakes on Self-Driving Cars Journal
40 Development of a Test Track for Driverless Cars: Vehicle Szalay, Z., Tettamanti, T., 2018 Periodica Polytechnica
Design, Track Configuration, and Liability Considerations Esztergár K. D., Varga, I. and Transportation
Bartolini, C Engineering
40 Splitting the Bill: Creating a National Car Insurance Fund Schroll, C 2015 Northwestern University
to Pay for Accidents in Autonomous Vehicles Law Review
36 The Ethical Knob: Ethically-Customisable Automated Contissa, G., Lagioia, F. and 2017 Artificial Intelligence and
Vehicles and the Law Sartor, G Law
35 Driving Miss Daisy: An Autonomous Chauffeur System Goodrich, J 2013 Houston Law Review

and improvement of AVs. In addition, some manufacturers case studies (i.e. survey or questionnaire), literature review
of AVs technology expect that personal ownership of AVs and content review (i.e. collecting, analyzing and com-
will be replaced by sharing AVs system, where companies paring existing knowledge to take advantage of new
own the AVs and they rent them to consumers who pay per requirements, laws and recommendations). Meanwhile, the
trip in these cases in case of an accident who hold liable articles provided recommendations to modify laws and
and how insurance should be regulated to pay for those regulations on liability and accountability issues. More-
accidents (i.e. drivers, car-sharing companies, or manu- over, how the public will react with the new law and
facturers).These articles used the following methodologies:

123
1242 Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249

liability. Also, these articles suggest a framework for which Developed countries such as the US, the UK and Ger-
courts could use when regulating liability in AV accidents. many are already moving in this direction, with working
Only a minority of articles used experimental testing and groups established to revise current laws and formulate
analysis for critical AV systems, such as programing, new ones. Policymakers are collaborating with other
sensing devices, data analysis, statistical risks and how stakeholders to ensure a fair legal system before AVs are
liability will affect AV manufacturers. Also, these articles adopted on a larger scale.
about decision informatics and contains the processing (i.e. The articles selected also discussed how the introduction
real-time analytics), advanced technologies of sensing (i.e. of AVs will shift liability from drivers to other stake-
Big Data), learning (i.e. deep learning) and reacting (i.e. holders. With driverless vehicles, occupants will no longer
real-time decision-making) which are qualitative in nature. be responsible for crashes: instead, the vehicle and its
These works mainly tested technical components of AVs systems will be responsible. Accordingly, automobile
that are related to technology manufacturers rather than companies and software companies that manufacture the
government entities and how liability will affect those vehicles, systems, devices and software for AVs will be
manufacturers. Also, how should AVs be equipped with liable for any accident. This shift will, in turn, change the
pre-programmed in case of unavoidable accident scenarios insurance industry and associated costs. Some researchers
and the liability issues in these cases. These articles rec- discussed the role of the insurance industry in ensuring an
ommended that manufacturers modify their systems to accident-free world.
make autonomous driving safe, secure, reliable and Whether the occupant will be completely spared from
affordable. all liabilities is also a topic of great interest. The occupants
and their ability to take any action to prevent an accident
are not ignored, indicating that despite the fascinating
4 Conclusions features of AVs, the importance of the human brain and its
efficiency in making timely decisions cannot be dismissed.
The present research conducted a systematic literature Researchers thus concluded that liability will depend on the
review on liability with AVs. AVs are the latest trend in the situation; no general judgment is made regarding the lia-
transportation industry, and they are changing the dynam- bility of AVs.
ics of the industry, including the liability of drivers when Despite the highly advanced technology involved in
there is a crash. AVs, risks and failures associated with safety and security
A detailed methodology was formulated to select the remain. In situations when humans can take over the
most relevant articles based on the inclusion and exclusion machine to make a timely decision to avoid an accident, the
criteria. Seventy-one articles were selected from three former should be made liable. This will keep vehicle
databases: EBSCO, SCOPUS and PROQUEST. All articles occupants motivated to avoid crashes (Pourgol-Mohammad
selected were peer reviewed and included the keywords et al. 2017). Researchers insist that liability should be
identified for this paper. The articles were then analyzed, collective, because putting the responsibility on only one
and the results were presented in this report. stakeholder will make other stakeholders lax. To ensure
Based on the findings, it can be concluded that research fair treatment, occupants, automobile companies and soft-
on liability related to AVs increased after 2015. Accord- ware companies are made liable, depending on the
ingly, the last five years have seen great achievements in situation.
the field of AVs, which in turn have encouraged Researchers also discussed the ethical aspects of AVs.
researchers to study liability issues in recent years. Most Researchers claim that the ethical responsibility of auto-
automobile companies have plans to launch fully auto- mobile and software companies should be identified by the
mated vehicles by late 2020, which will change the liability government and by policymakers, and clear and open
aspect of driving. disclosure of information should be mandatory. Manufac-
The majority of the articles were published in law turers should be transparent in the testing of AVs before
journals, and most researchers discussed the legal compo- market launch.
nents of the issue. These articles investigate how the mass Since the successful implementation of AVs involves all
adoption of AVs will change the legal environment of the stakeholders working closely with automobile and software
countries involved and how the industry is taking a keen companies should act ethically and ensure the free flow of
interest in facilitating that change. Researchers affirm that information. Hiding sensitive information for profit and
current laws and regulations will not be applicable after the monetary gain should be strictly prohibited, since any
introduction of AVs. Current laws need to be amended, and failure can lead to accidents, chaos and loss of human lives.
new laws need to be implemented to ensure a regulatory Similarly, third-party evaluators should be encouraged to
framework appropriate for automated driving.

123
Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249 1243

give honest reviews of AVs and their testing and should research. Further research is needed to study particular
avoid bias and monetary incentives from companies. countries and users with various profiles and to examine
From the critical evaluation of the selected articles, it how users’ perceptions of liability will change and how
can be concluded that most articles on liability with AVs users will react to newly formulated laws.
concentrate on qualitative approaches and recommend
changes to laws and regulations for autonomous driving. A
minority of articles deal with technological experiments on Appendix
autonomous systems and recommend amendments for
manufacturers and technology companies. This paper See Table 7.
affirms that AV-related liability is an ongoing topic in

Table 7 List of articles reviewed


No. Authors Title Year Journal

1 Markham, T. R., & Chernoguzov, A A Balanced Approach For Securing The OBD-II Port 2017 SAE International Journal
of Passenger Cars
2 Murphy, R., Shell, D., Guerin, A., A Midsummer Night’s Dream (With Flying Robots) 2011 Autonomous Robots
Duncan, B., Fine, B., Pratt, K., &
Zourntos, T
3 Martinesco, A., Netto, M., Neto, A. M., A Note On Accidents Involving Autonomous Vehicles: 2019 IFAC-PapersOnLine
& Etgens, V. H Interdependence Of Event Data Recorder, Human-
Vehicle Cooperation And Legal Aspects
4 Geistfeld, M. A A Roadmap For Autonomous Vehicles: State Tort 2017 California Law Review
Liability, Automobile Insurance, And Federal Safety
Regulation
5 McCall, R., McGee, F., Mirnig, A., A Taxonomy Of Autonomous Vehicle Handover 2019 Transportation Research
Meschtscherjakov, A., Louveton, N., Situations Part A: Policy and
Engel, T., & Tscheligi, M Practice
6 Casey, B Amoral Machines, Or: How Roboticists Can Learn To 2017 Northwestern University
Stop Worrying And Love The Law Law Review
7 Mat, A. R., How, L. M., Ariff, O. K., Autonomous Aerial Hard Docking Of Fixed And Rotary 2014 Applied Mechanics and
Zhahir, M. A. M., & Ajir, R. M. Wing Uavs: Task Assessment And Solution Architecture Materials
8 de Bruin, R Autonomous Intelligent Cars On The European 2016 European Journal of Risk
Intersection Of Liability And Privacy Regulation: EJRR
9 Brodsky, J. S Autonomous Vehicle Regulation: How An Uncertain 2016 Berkeley Technology Law
Legal Landscape May Hit The Brakes On Self-Driving Journal
Cars
10 Olalekan, O. S Autonomous Vehicles And Tortious Liability: An Islamic 2018 Jurnal Syariah
Perspective
11 Zhang, Z., Singh, J., Gadiraju U., & Dissonance Between Human And Machine Understanding 2019 Proceedings Of The ACM
Anand, A On Human–Computer
Interaction
12 Lemann, A. B Autonomous Vehicles, Technological Progress, And The 2019 Journal Of Tort Law
Scope Problem In Products Liability
13 Gandia, R. M., Antonialli, F., Cavazza, Autonomous Vehicles: Scientometric And Bibliometric 2019 Transport Reviews
B. H., Neto, A. M., Lima, D. A. D., Review
Sugano, J. Y.,
& Zambalde, A. L
14 Sinclair J., Schafer B Autonomous Vehicles: The Path To Liability Is Still 2017 Jusletter IT
Unclear
15 Hammond, D. N Autonomous Weapons And The Problem Of State 2015 Chicago Journal of
Accountability International Law
16 Hopkins, J., & Hawking, P Big Data Analytics And Iot In Logistics: A Case Study 2018 International Journal of
Logistics Management
17 Calo, R Commuting To Mars: A Response To Professors Abraham 2019 Virginia Law Review
And Rabin

123
1244 Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249

Table 7 continued
No. Authors Title Year Journal

18 Correa, O., Khan, A. M. R., Tanin, E., Congestion-Aware Ride-Sharing 2019 ACM Transactions on
Kulik, L., & Ramamohanarao, K Spatial Algorithms and
Systems (TSAS)
19 Sheehan, B., Murphy, F., Mullins, M., & Connected And Autonomous Vehicles: A Cyber-Risk 2019 Transportation Research
Ryan, C Classification Framework Part A: Policy and
Practice
20 Mrcela, M., & Vuletic, I Criminal Law Facing Challenges Of Robotics: Who Is 2018 Zbornik Pravnog fakulteta
Liable For Traffic Accident Caused By Autonomous u Zagrebu
Vehicle? [Kazneno Pravo Pred Izazovima Robotike:
Tko Je Odgovoran Za Prometnu Nesreću Koju Je
Prouzročilo Neovisno Vozilo?]
21 Liivak, T., & Lahe, J Delictual Liability For Damage Caused By Fully 2018 Masaryk University
Autonomous Vehicles: The Estonian Perspective Journal of Law and
Technology
22 Szalay, Z., Tettamanti, T., Esztergár- Development Of A Test Track For Driverless Cars: 2018 Periodica Polytechnica.
Kiss, D., Varga, I., & Bartolini, C Vehicle Design, Track Configuration, And Liability Transportation
Considerations Engineering
23 Danas, A. M Disruptive Technologies And Business Models: Emerging 2018 Journal of
Regulatory Issues From The Sharing Economy Multidisciplinary
Research
24 Genziuk, Shane Don’t Blame Me—Blame The Car 2015 General Insurance
25 Jones L Driverless Cars: When And Where? 2017 Engineering and
Technology
26 Goodrich, J Driving Miss Daisy: An Autonomous Chauffeur System 2013 Houston Law Review
27 Aguilar, A. B Eco-Taxes In Spain 2000 European Energy and
Environmental Law
Review
28 Murphy, F., & Mullins, M Editorial For The Special Issue – Liability And Insurance 2019 Transportation Research
For Semi-Autonomous Vehicles Part A: Policy and
Practice
29 Shabanpour, R., Golshani, N., Eliciting Preferences For Adoption Of Fully Automated 2019 Transportation research
Shamshiripour, A., & Mohammadian, Vehicles Using Best–Worst Analysis part C: emerging
A. K technologies
30 Dhar, V Equity, Safety, And Privacy In The Autonomous Vehicle 2016 IEEE Computer Society
Era
31 Shadrin, S. S., Varlamov, O. O., & Experimental Autonomous Road Vehicle With Logical 2017 Journal of Advanced
Ivanov, A. M Artificial Intelligence Transportation
32 Lederman, J., Garrett, M., & Taylor, B. Faulty Reasoning: Navigating The Liability Terrain In 2016 Public Works
D Intelligent Transportation Systems Management and Policy
33 Humaidi, A. J., Hassan, S., & Fadhel, M. FPGA-Based Lane-Detection Architecture For 2018 Asia Life Sciences
A Autonomous Vehicles: A Real-Time Design And
Development
34 Bellet, T., Cunneen, M., Mullins, M., From Semi To Fully Autonomous Vehicles: New 2019 Transportation research
Murphy, F., Pütz, F., Spickermann, F., Emerging Risks And Ethico-Legal Challenges For part F: traffic
& Baumann, M. F Human–Machine Interactions psychology and
behaviour
35 Taeihagh, A., & Lim, H. S. M Governing Autonomous Vehicles: Emerging Responses 2018 Transport Reviews
For Safety, Liability, Privacy, Cybersecurity, And
Industry Risks
36 Lucas Jr, G. R Industrial Challenges Of Military Robotics 2012 Journal of Military Ethica
37 Fan, C. K., & Xu, X Influences Of Autonomous Cars On The Insurance Market 2019 Journal of Applied
From The Perspectives Of Insurance Companies And Finance and Banking
Auto Insurance Agencies
38 Thomas, Mike INSURANCE: Challenges To The Business Model: The 2017 JASSA
Australasian Journal Of Applied Finance

123
Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249 1245

Table 7 continued
No. Authors Title Year Journal

39 Burri, T., & Wildhaber, I Introduction To The Special Issue On The Man And The 2016 European Journal of Risk
Machine Regulation: EJRR
40 Gauci, G. M Is It A Vessel, A Ship Or A Boat, Is It Just A Craft, Or Is It 2016 Journal of Maritime Law
Merely A Contrivance? & Commerce
41 Juhász, Á., & Pusztahelyi, R Legal Questions On The Appearance Of Self-Driving Cars 2016 European Integration
In The Road Traffic With Special Regard On The Civil Studies
Law Liability
42 O’Sullivan, S., Nevejans, N., Allen, C., Legal, Regulatory, And Ethical Frameworks For 2019 International Journal of
Blyth, A., Leonard, S., Pagallo, U., & Development Of Standards In Artificial Intelligence (AI) Medical Robots
Ashrafian, H And Autonomous Robotic Surgery
43 De Bruyne, J., & Tanghe, J Liability For Damage Caused By Autonomous Vehicles: 2018 Journal of European Tort
A Belgian Perspective Law
44 Pinter, K., Szalay, Z., & Vida, G Liability In Autonomous Vehicle Accidents 2017 Komunikácie
45 Lohmann, M. F Liability Issues Concerning Self-Driving Vehicles 2016 European Journal of Risk
Regulation: EJRR
46 De Bruyne, J., & Werbrouck, J Merging Self-Driving Cars With The Law 2018 Computer Law and
Security Review
47 Schellekens, M No-Fault Compensation Schemes For Self-Driving 2018 Law, Innovation and
Vehicles Technology
48 Hoffmann, T., & Prause, G On The Regulatory Framework For Last-Mile Delivery 2018 Machines
Robots
49 Determann, L., & Perens, B Open Cars 2017 Berkeley Technology law
Journal
50 Hall, S P-71 Evaluation Of A Rapid Discharge Pathway For 2017 BMJ Journals
Dying Patients
51 Tscharaktschiew, S., & Evangelinos, C Pigouvian Road Congestion Pricing Under Autonomous 2019 Transportation Research
Driving Mode Choice Part C: Emerging
Technologies
52 Fagnant, D. J., & Kockelman, K Preparing A Nation For Autonomous Vehicles: 2015 Transportation Research
Opportunities, Barriers And Policy Recommendations Part A: Policy and
Practice
53 Collingwood, L Privacy Implications And Liability Issues Of Autonomous 2017 Information and
Vehicles Communications
Technology Law
54 Fabian, P. Ü. T. Z., Murphy, F., Mullins, Reasonable, Adequate And Efficient Allocation Of 2018 European Journal of Risk
M., Maier, K., Friel, R., & Rohlfs, T Liability Costs For Automated Vehicles: A Case Study Regulation: EJRR
Of The German Liability And Insurance Framework
55 Hevelke, A., & Nida-Rümelin, J Responsibility For Crashes Of Autonomous Vehicles: An 2015 Science and Engineering
Ethical Analysis Ethics
56 Wang, S., & Zhao, J Risk Preference And Adoption Of Autonomous Vehicles 2019 Transportation Research
Part A: Policy and
Practice
57 Bogue R Robot Ethics And Law 2014 Industrial Robot
58 Schellekens, M Self-Driving Cars And The Chilling Effect Of Liability 2015 Computer Law and
Law Security Review
59 Sheehan, B., Murphy, F., Ryan, C., Semi-Autonomous Vehicle Motor Insurance: A Bayesian 2017 Transportation Research
Mullins, M., & Liu, H. Y Network Risk Transfer Approach Part C: Emerging
Technologies
60 Boothby, W Some Legal Challenges Posed By Remote Attack 2012 International Review of
the Red Cross
61 Schroll, C Splitting The Bill: Creating A National Car Insurance 2015 Northwestern University
Fund To Pay For Accidents In Autonomous Vehicles Law Review
62 Rosenberg, A Strict Liability: Imagining A Legal Framework For 2015 Tulane Journal of
Autonomous Vehicles Technology &
Intellectual Property

123
1246 Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249

Table 7 continued
No. Authors Title Year Journal

63 Brock, L., & Tropnas, L Survey On The Regulations Of Autonomous Vehicles 2018 Journal of
Multidisciplinary
Research
64 Contissa, G., Lagioia, F., & Sartor, G The Ethical Knob: Ethically-Customisable Automated 2017 Artificial Intelligence and
Vehicles And The Law Law
65 Ryan, M The Future Of Transportation: Ethical, Legal, Social And 2019 Science and Engineering
Economic Impacts Of Self-Driving Vehicles In The Ethics
Year 2025
66 Showalter, S The Growing Use Of Auvs And Liability: Should 2006 International Ocean
Operators Be Concerned? Systems
67 Churilov, A. Y The Legal Basis Of Liability For The Damage Caused 2018 Legal Concept
During The Operation Of An Autonomous Vehicle
68 Zipp, J. W The Road Will Never Be The Same: A Reexamination Of 2016 Transportation Law
Tort Liability For Autonomous Vehicles Journal
69 Tien, J. M The Sputnik Of Servgoods: Autonomous Vehicles 2017 Journal of Systems
Science and Systems
Engineering
70 Mangan, M The Value Of Vanpooling As A Strategic, Cost-Effective, 2018 ITE Journal
And Sustainable Transportation Option
71 Beale, A. F Who’s Coffers Spill When Autonomous Cars Kill? A New 2018 Widener Law Journal
Tort Theory For The Computer Code Road

References Baglietto E, Consilvio A, Di Febbraro A, Papa F, Sacco N (2018) A


Bayesian network approach for the reliability analysis of
Abbott R (2018) The reasonable computer: disrupting the paradigm of complex railway systems. In: 2018 International Conference on
tort liability. Geo Wash L Rev 86:1 Intelligent Rail Transportation (ICIRT), pp. 1–6
Abdullah Z (2016) Driverless vehicles could change laws, insurance Barfield W (2018) Liability for autonomous and artificially intelligent
policies. The straits times. Retrieved from https://www.strait robots. Paladyn J Behav Robot 9(1):193–203
stimes.com/singapore/transport/driverless-vehicles-could- Bartolini C, Tettamanti T, Varga I (2017) Critical features of
change-laws-insurance-policies autonomous road transport from the perspective of technological
Abraham K (2017) The forms and functions of tort law. West regulation and law. Transport Res Proc 27:791–798
Baumann MF, Brändle C, Coenen C, Zimmer-Merkle S (2019)
Academic, Saint Paul
Abraham KS, Rabin RL (2019a) Automated vehicles and manufac- Taking responsibility: a responsible research and innovation
turer responsibility for accidents: a new legal regime for a new (RRI) perspective on insurance issues of semi-autonomous
era. Va L Rev 105:127 driving. Transport Res Part A Policy Pract 124:557–572
Beiker S (2016) Deployment scenarios for vehicles with higher-order
Abraham KS, Rabin RL (2019b) The future is almost here: inaction is
actually mistaken action. Va L Rev Online 105:91 automation. Autonomous driving. Springer, Berlin, pp 193–211
Ackerman RP (2019) Regulating autonomous vehicles in the United Boeglin J (2015) The costs of self-driving cars: reconciling freedom
States. and privacy with tort liability in autonomous vehicle regulation.
Ahmed R (2019) The influence of ’reasonableness’ on the element of Yale JL Tech 17:171
conduct in delictual or tort liability-comparative conclusions. Bomey N (2018) Tesla model X driver killed in California Crash
Potchefstroom Electron Law J 22 wasn’t Holding Steering Wheel, NTSB says.’’ USA Today. June
7th, 2018. Accessed January 22nd, 2019. https://www.usatoday.
Aleksandrovskaya LN, Ardalionova AE, Papic L (2019) Application
of probability distributions mixture of safety indicator in risk com/story/money/cars/2018/06/07/tesla-model-x-autopilot-
assessment problems. Int J Syst Assur Eng Manag 10(1):3–11 crash-ntsb-report/681148002/
Alessandrini A, Campagna A, Delle Site P, Filippi F, Persia L (2015) Bramer WM, Rethlefsen ML, Kleijnen J, Franco OH (2017) Optimal
database combinations for literature searches in systematic
Automated vehicles and the rethinking of mobility and cities.
Transport Res Proc 5(2015):145–160 reviews: a prospective exploratory study. Syst Rev 6:245
Attias, D. (2017). The autonomous car, a disruptive business model?. Browning JG (2014) Emerging technology and its impact on
In The Automobile Revolution (pp. 99–113). Springer, Cham. automotive litigation. Defense Counsel J 81(1):83–90
Australian Government (2017) Social Impacts of Automation in Casey B (2016) A loophole large enough to drive an autonomous
Transport, Department of Infrastructure and Regional vehicle through: the ADA’s new van provision and the future of
Development access to transportation. Stan L Rev Online 69:73
Bacca J, Baldiris S, Fabregat R, Graf S, Kinshuk. (2014) Augmented CCAV (2016) Pathway to driverless cars: proposals to support
reality trends in education: a systematic review of research and advanced driver assistance systems and automated vehicle
applications. J Educ Technol Soc 17(4):133 technologies, Centre for connected and autonomous vehicles.
Available at: https://www.gov.uk/government/uploads/system/

123
Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249 1247

uploads/attachment_data/file/536365/driverless-cars-proposals- obligation to insure against such liability’, Official Journal of the


for-adas-and_avts.pdf European Union, pp. 11–31.
Chatzipanagiotis MP, Leloudas G (2020) Automated vehicles and European Union (2016) Declaration of Amsterdam: cooperation in the
third-party liability: A European perspective. University of field of connected and automated driving, The Netherlands EU
Illinois Journal of Law, Technology & Policy, Forthcoming. Presidency. Available at: https://english.eu2016.nl/binaries/
Choi BH (2019) Crashworthy code. Wash L Rev 94:39 eu2016-en/documents/publications/2016/04/14/declaration-of-
CNA (2017) Regulations in place to ramp up driverless vehicle trials amsterdam/2016-04-08-declaration-of-amsterdam-final-format-
in Singapore. Channel News Asia. 3.pdf
Cohen T, Jones P, Cavoli C (2017) Social and behavioural questions Faisal A, Yigitcanlar T, Kamruzzaman M, Currie G (2019) Under-
associated with automated vehicles. Scoping study by UCL standing autonomous vehicles: a systematic literature re-view on
Transport Institute. Final report, London: Department for capability, impact, planning and policy. J Transp Land Use
Transport. 12:45–72
Collingwood L (2017) Privacy implications and liability issues of Fagnant DJ, Kockelman K (2015) Preparing a nation for autonomous
autonomous vehicles. Inform Commun Technol Law vehicles: opportunities, barriers and policy recommendations.
26(1):32–45 Transport Res Part A Policy Pract 77:167–181
Consilvio A, Di Febbraro A, Sacco N, Torre A (2019) On exploring Fleetwood J (2017) Public health, ethics, and autonomous vehicles.
the potentialities of autonomous vehicles in urban spatial Am J Public Health 107(4):532–537
planning. In: 2019 6th International Conference on models and Freemark Y, Hudson A, Zhao J (2019) Are cities prepared for
technologies for intelligent transportation systems (MT- autonomous vehicles? Planning for technological change by US
ITS) (pp. 1–7). IEEE local governments. J Am Plann Assoc 85(2):133–151
Consilvio A, Sanetti P, Anguı̀ta D, Crovetto C, Dambra C, Oneto L, Geistfeld MA (2017) A roadmap for autonomous vehicles: State tort
Sacco N (2019) Prescriptive maintenance of railway infrastruc- liability, automobile insurance, and federal safety regulation.
ture: from data analytics to decision support. In :2019 6th Calif L Rev 105:1611
International Conference on Models and Technologies for Genziuk S (2015) Don’t blame me—blame the car. J the Aust N Zeal
Intelligent Transportation Systems (MT-ITS) (pp. 1–10). IEEE. Inst Insur Financ 38(2):1–3
Cowger AR Jr (2018) Liability considerations when autonomous Gurney J (2016) Imputing driverhood: applying a reasonable driver
vehicles choose the accident victim. J High Tech L 19:1 standard to accidents caused by autonomous vehicles. Forth-
Dai S (2018) China’s Google’’ says government support and coming, Robot Ethics, 2
population size matters in global AI race. South China Morning Gurney JK (2013) Sue my car not me: Products liability and accidents
Post. involving autonomous vehicles. U. Ill. JL Tech. & Pol’y, 247
Daisuke W (2018) Self-driving uber kills pedestrian in Arizona, Hamidi S, Ewing R, Preuss I, Dodds A (2015) Measuring sprawl and
Where Robots Roam. The New York Times. March 19th, 2018. its impacts: an update. J Plan Educ Res 35(1):35–50
Accessed January 31st, 2019. https://www.nytimes.com/2018/ Hawkins A (2018) Uber Self-Driving car saw pedestrian but didn’t
03/19/technology/uber-driverless-fatality.html brake before fatal crash, feds say. The Verge. May 24th, 2018.
Di Fabio U, Broy M, Brüngger RJ (2017) Ethics commission Accessed February 7th, 2018. https://www.theverge.com/2018/5/
automated and connected driving. Federal Ministry of Transport 24/17388696/uber-self-driving-crash-ntsb-report
and Digital Infrastructure of the Federal Republic of Germany HC143 (2017) Vehicle technology and aviation bill HC Bill 143.
Dhar V (2016) Equity, safety, and privacy in the autonomous vehicle Hevelke A, Nida-Rümelin J (2015) Responsibility for crashes of
era. Computer 49(11):80–83 autonomous vehicles: an ethical analysis. Sci Eng Ethics
Douma F, Palodichuk SA (2012) Criminal liability issues created by 21(3):619–630
autonomous vehicles. Santa Clara L Rev 52:1157 James AT, Gandhi OP, Deshmukh SG (2018) Fault diagnosis of
Duarte F, Ratti C (2018) The impact of autonomous vehicles on automobile systems using fault tree based on digraph modeling.
cities: a review. J Urban Technol 25(4):3–18 Int J Syst Assur Eng Manag 9(2):494–508
Duffy SH, Hopkins JP (2013) Sit, stay, drive: the future of Japan Bullet (2017) New guidelines to allow driverless vehicle tests
autonomous car liability. SMU Sci Tech L Rev 16:453 on public roads. x: Japan Bullet
Eastman AD (2016) Is no-fault auto insurance the answer to liability JDSUPRA (2017) Germany permits automated vehicles. Available at:
concerns of autonomous vehicles. Am J Bus Manag 5(3):85–90 https://www.whitecase.com/publications/article/germany-per
Ebnali M, Hulme K, Ebnali-Heidari A, Mazloumi A (2019) How does mits-automated-vehicles
training effect users’ attitudes and skills needed for highly Jiji (2017) Japan sets approval criteria for driverless vehicle road
automated driving? Transport Res Part F 66:184–195 tests. The Japan Times
European Parliament (2017) Robots and artificial intelligence: MEPs Johnson MP (2001) Environmental impacts of urban sprawl: a survey
call for EU-wide liability rules. European Parliament. Available of the literature and proposed research agenda. Environ Plann A
at: https://www.europarl.europa.eu/news/en/press-room/ 33(4):717–735
20170210IPR61808/robots-and-artificial-intelligence-meps-call- Kalra N, Paddock SM (2016) Driving to safety: How many miles of
for-eu-wide-liability-rules driving would it take to demonstrate autonomous vehicle
EPCLA (2016) Draft report with recommendations to the Commis- reliability? Transport Res Part A Policy Pract 94:182–193
sion on Civil Law Rules on Robotics (2015/2103(INL), Euro- Kang N, Feinberg FM, Papalambros PY (2017) Autonomous electric
pean Parliament Committee on Legal Affairs vehicle sharing system design. J Mech Design 139(1)
European Parliament (2016) Policy department B: structural and Kerr D (2018) Are driverless cars safe? Uber fatality raises questions.
cohesion policies. https://www.europarl.europa.eu/RegData/ C|Net. March, 21st, 2018. Accessed February 3rd, 2019. https://
etudes/STUD/2016/573434/IPOL_STU(2016)573434_EN.pdf www.cnet.com/news/uber-self-driving-car-fatality-in-arizona-
European Union (2009) ‘DIRECTIVE 2009/103/EC OF THE has-people-asking-how-safe-are-driverless-cars/
EUROPEAN PARLIAMENT AND OF THE COUNCIL of 16 Ki M (2015) Driver at the wheel? Self-driving vehicles and the traffic
September 2009 relating to insurance against civil liability in and transport system of the future. Ministry of Infrastructure and
respect of the use of motor vehicles, and the enforcement of the the Environment, Hague
Kirkpatrick K (2013) Legal issues with robots

123
1248 Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249

Kohn LT, Corrigan J, Donaldson MS (2000) To err is human: National Transportation Safety Board (2018b) Preliminary Report
building a safer health system, vol 6. National academy press, Highway: HWY18MH010’’. May 24th, 2018. Accessed January
Washington 23rd, 2019. https://www.ntsb.gov/investigations/accidentreports/
Koopman P, Wagner M (2016) Challenges in autonomous vehicle pages/hwy18mh010-prelim.aspx
testing and validation. SAE Int J Transport Saf 4(1):15–24 National Transportation Safety Board (2018c) Preliminary Report
Krisher T (2018) Uber Self-driving SUV saw pedestrian but did not Highway: HWY18FH0111. June 7th, 2018. Accessed January
brake, federal report finds. Chicago Tribune. May 24th, 2018. 23rd, 2019. https://www.ntsb.gov/news/press-releases/pages/
Accessed February 7th, 2019. https://www.chicagotribune.com/ nr20180607.aspx
business/ct-biz-uber-self-driving-uber-car-report-20180524- National Transportation Safety Board (2017) Driver errors, overre-
story.html liance on automation, lack of safeguards, led to fatal tesla crash.
Kyriakidis M, Happee R, Winter JC (2015) Public opinion on September 12, 2017. Accessed January 25th, 2019. https://www.
automated driving: results of an international questionnaire ntsb.gov/news/press-releases/Pages/PR20170912.aspx
among 5000 respondents transport. Res Part F Traffic Psychol NHTSA (2017) Automated Driving Systems 2.0 a vision for safety,
Behav 32(2015):127–140 National Highway Traffic Safety Administration, U.S. Depart-
Lam RW, Kennedy SH (2005) Using metaanalysis to evaluate ment of Transportation
evidence: practical tips and traps. Can J Psychiat 50:167–174 Nikitas A, Njoya ET, Dani S (2019) Examining the myths of
Laris M (2018) Self-driving uber did not have emergency braking connected and autonomous vehicles: analysing the pathway to a
turned on when it hit pedestrian, NTSB says. The Washington driverless mobility paradigm. Int J Automot Technol Manage
Post. May 24th, 2018. Accessed January 14th, 2019. https:// 19(1–2):10–30
www.washingtonpost.com/news/dr-gridlock/wp/2018/05/24/ Nikkei (2018) Japan eyes black boxes for automated vehicles. Nikkei
ntsb-self-driving-uber-did-not-have-emergency-braking-turned- Asian Rev
on/? Niles J (2019) Full potential of future robotaxis achievable with trip-
Lederman J, Garrett M, Taylor BD (2016) Fault-y reasoning: based subsidies and fees applied to the for-hire vehicles of today.
navigating the liability terrain in intelligent transportation Mineta Transportation Institute Publications
systems. Public Works Manag Policy 21(1):5–27 NTC (2017) Standing Committee on Industry, Innovation, Science
Lohmann M (2016) Liability issues concerning self-driving vehicles. and Resources’ Inquiry into the Social Issues Relating to Land-
Eur J Risk Regul 7(2):335–340 Based Driverless Vehicles in Australia. Available at: https://
Madadi B, van Nes R, Snelder M, van Arem B (2020) A bi-level www.aph.gov.au/DocumentStore.ashx?id=10e60133-f61e-44e5-
model to optimize road networks for a mixture of manual and 9365-2d2928b3a129&subId=509089
automated driving: an evolutionary local search algorithm. Patti FP (2019) The European road to autonomous vehicles. Fordham
Comput Aided Civ Infrastruct Eng 35(1):80–96 Int’l LJ 43:125
Mallett R, Hagen-Zanker J, Duvendack M, Slater R (2012) The Pendleton SD, Andersen H, Du X, Shen X, Meghjani M, Eng YH,
benefits and challenges of using systematic reviews in interna- Ang MH (2017) Perception, planning, control, and coordination
tional development research. J Dev Effect 4(3):445–455 for autonomous vehicles. Machines 5:6
Marçal R, Braga R, Antonialli F, Cavazza BH, Sugano JY, Castro C, Perc M, Ozer M, Hojnik J (2019) Social and juristic challenges of
Nicolaı̈ I (2017) The quintuple helix model and the future of artificial intelligence. Palgrave Commun 5(1):1–7
mobility: the case of autonomous vehicles. In: 25th International Porter L, Stone J, Legacy C, Curtis C, Harris J, Fishman E, Stilgoe J
Colloquium of Gerpisa-R/Evolutions. New Technologies and (2018) The autonomous vehicle revolution: implications for
Services in the Automotive Industry, June, Paris, France planning/the driverless city?/Autonomous vehicles–a planner’s
Marchant GE, Lindor RA (2012) The coming collision between response/autonomous vehicles: opportunities, challenges and the
autonomous vehicles and the liability system. Santa Clara L Rev need for government action/three signs autonomous vehicles will
52:1321 not lead to less car ownership and less car use in car dependent
Marsden G, Reardon L (2017) Questions of governance: rethinking cities–a case study of Sydney, Australia/Planning for Autono-
the study of transportation policy. Transport Res Part A Policy mous Vehicles? Questions of purpose, place and pace/ensuring
Pract 101:238–251 good governance: the role of planners in the development of
McChristian L, Corbett R (2016) Regulatory issues related to autonomous vehicles. Plann Theory Pract 19(5):753–778
autonomous vehicles. J Insur Regul 35(7) Pourgol-Mohammad M, Hejazi A, Soleimani M, Ghasemi P, Ahmadi
Merat N, Jamson AH, Lai FC, Daly M, Carsten OM (2014) Transition A, Jalali-Vahid D (2017) Design for reliability of automotive
to manual: driver behaviour when resuming control from a systems; case study of dry friction clutch. Int J Syst Assur Eng
highly automated vehicle. Transport Res Part F 27:274–282 Manag 8(3):572–583
Mladenovic MN, McPherson T (2016) Engineering social justice into Powell JP, Fraszczyk A, Cheong CN, Yeung HK (2016) Potential
traffic control for self-driving vehicles? Sci Eng Ethics benefits and obstacles of implementing driverless train operation
22(4):1131–1149 on the Tyne and Wear Metro: a simulation exercise. Urban Rail
MOT (2017) Opening speech by second minister for transport Ng Transit 2(3–4):114–127
Chee Meng for the road traffic (amendment) bill second reading. Ramirez E (2017) How South Korea plans to put driverless cars on
Ministry of Transport, Singapore the road by 2020. Forbes.
National Conference of State Legislatures (2016) Autonomous self- Riehl DA (2018) Car minus driver: autonomous vehicles driving
driving vehicles legislation. https://www.ncsl.org/research/trans regulation, liability, and policy. Comput Internet Lawyer
portation/autonomous-vehicles-legislation.aspx. Accessed 02 35(5):1–18
Nov 2016 Ross PE (2014) Robot, you can drive my car; autonomous driving
National Transportation Safety Board (2018a) Preliminary report will push humans into the passenger seat. IEEE Spectr
highway HWI18FH011. National Transportation Safety Board, 51(6):60–90
1–4. March 23rd, 2018. Accessed January 28th, 2019. https:// Ryan C (2017) Driverless cars: preparing for the impact on the
ntsb.gov/investigations/AccidentReports/Reports/ automobile insurance industry
HWY18FH011-preliminary.pdf SAE On-Road Automated Vehicle Standards Committee (2018)
Taxonomy and definitions for terms related to driving

123
Int J Syst Assur Eng Manag (December 2020) 11(6):1227–1249 1249

automation systems for on-road motor vehicles. SAE Interna- Stone T, de Sio FS, Vermaas PE (2020) Driving in the dark: designing
tional, Warrendale autonomous vehicles for reducing light pollution. Sci Eng Ethics
Schellekens M (2015) Self-driving cars and the chilling effect of 26(1):387–403
liability law. Comput Law Secur Rev 31(4):506–517 Sun Y, Olaru D, Smith B, Greaves S, Collins A (2016) Road to
Schoitsch E (2016) Autonomous vehicles and automated driving autonomous vehicles in Australia: a comparative literature
status, perspectives and societal impact. Information technology, review. In: Australasian transport research forum, pp. 16–18
society and economy strategic cross-influences (IDIMT-2016). Suresh P, Manivannan PV (2014) Reduction of vehicular pollution
24th Interdisciplinary Information Management Talks, 45(1): through fuel economy improvement with the use of autonomous
405–424. self-driving passenger cars. J Environ Res Dev 8(3A):705
Schreurs MA, Steuwer SD (2015) Autonomous driving-political, Taeihagh A, Lim HSM (2018) Governing autonomous vehicles:
legal, social, and sustainability dimensions. In: Autonomes emerging responses for safety, liability, privacy, cybersecurity,
Fahren. Springer Vieweg, Berlin, pp. 151–173 and industry risks. Transp Rev 39(1):103–128
Selbst AD (2019) Negligence and AI’s human users. Boston Tien JM (2017) The Sputnik of servgoods: autonomous vehicles.
University Law Review, Forthcoming. J Syst Sci Syst Eng 26(2):133–162
Shladover SE (2018) Connected and automated vehicle systems. UNECE (2012) Design principles for control systems of ADAS.
J Intell Transport Syst 22:190–200 https://www.unece.org/fileadmin/DAM/trans/doc/2011/wp29/
Shaheen SA, Mallery MA, Kingsley KJ (2012) Personal vehicle ITS-19-07e.pdf
sharing services in North America. Res Transport Bus Manag Urmson C (2016) Google self-driving car project, paper presented at
3:71–81 the South by 31 Southwest (SXSW) 2016. Austin, Texas
Shavell S (2019) On the redesign of accident liability for the world of Urooj S, Feroz I, Ahmad N (2018) Systematic literature review on
autonomous vehicles (No. w26220). National Bureau of Eco- user interfaces of autonomous cars: liabilities and responsibil-
nomic Research ities. In: 2018 International Conference on Advancements in
Singh S (2015) Critical reasons for crashes investigated in the Computational Sciences (ICACS) (pp. 1–10). IEEE
national motor vehicle crash causation survey (No. DOT HS 812 Vleugel JM, Bal F (2018) More space and improved living conditions
115) in cities with autonomous vehicles. Int J Des Nat Ecodyn
Smith BW (2016) Automated driving and product liability’. Detroit 12(4):505–515
College Law Michigan State Univ Law Rev 1:1 Wacket M, Escritt T, Davis T (2017) Germany adopts self-driving
Soo Y (2018) Utah driver who slammed her Tesla into a fire truck vehicles law. Reuters.
sues the carmaker over Autopilot feature. ABC News, September Walpert JD (2016) Carpooling liability: applying tort law principles
6th, 2018. Accessed January 31st, 2019. https://abcnews.go.com/ to the joint emergence of self-driving automobiles and trans-
beta-story-container/Technology/utah-driver-slammed-tesla-fire portation network companies. Fordham L Rev 85:1863
truck-sues-carmaker-autopilot/story Webb KC (2016) Products liability and autonomous vehicles: who’s
Sparrow R, Howard M (2017) When human beings are like drunk driving whom. Rich JL & Tech 23:1
robots: driverless vehicles, ethics, and the future of transport. West DM (2016) Moving forward: self-driving vehicles in China,
Transport Res Part C Emerg Technol 80:206–215 Europe, Japan, Korea, and the United States. Washington, DC,
State of California (2020) Report of traffic collison involving an USA, Center for Technology Innovation at Brookings
autonomous vehicle (OL 316). State of California, Department Woods, J. B. (2019). Autonomous Vehicles and Police De-Escalation.
of Motor Vehicles. Accessed January 31st, 2020. https://www. Zakharenko R (2016) Self-driving cars will change cities. Region Sci
dmv.ca.gov/portal/dmv/detail/vr/autonomous/autonomousveh_ Urban Econ 61:26–37
ol316?
Stocker A, Shaheen S (2017) Shared automated vehicles: review of
Publisher’s Note Springer Nature remains neutral with regard to
business models. International Transport Forum Discussion
jurisdictional claims in published maps and institutional affiliations.
Paper

123

You might also like