Legalities of Autonomous Vehicles

By Jiya Bharti

During my final year of Law school, I wrote a brief essay on autonomous vehicles. Since then, I have revised it and decided to share it due to its ongoing relevance. 

The industry based upon autonomous vehicles has an estimated 16% increase globally each year. [1] Under UK law, autonomous vehicles fall under the jurisdiction S1 Automated and Electric Vehicles Act 2018, where they are defined as vehicles capable to safely drive themselves in some situations. The role of law in relation to autonomous vehicles is arguably undetermined, since autonomous vehicles as technology is a new and constantly developing concept. Some legal issues pertaining to the development, the liability, the financial realm and also the very absence of law itself, could arise. Furthermore, there are also ethical concerns, including the trolley problem, which raises concerns about discrimination and decisions made by autonomous vehicles. There are concerns about how developers can programme ethical rules into cars, and resultant ethical concerns about how these rules should be determined. 

Data Protection and Privacy

Autonomous vehicles would rely on networks and data, therefore, there is the question of privacy perception based on this new technology. They would have to constantly capture information on the environment and the user, which would create very little privacy. This is why policies would have to be thorough and constantly investigated. One privacy concern would be the invasion of privacy through location data. Autonomous vehicles would have records of precise locations and although this would use data to help improve safety and traffic planning, it would be recognised as a privacy concern. A case which can be referenced in relation to location privacy would be that of United States v. Jones, 565 U.S. 400 (2012), in which the Supreme Court acknowledged that the data collected on location has a vast impact on privacy as it “generates a precise, comprehensive record of a person’s public movements that reflects a wealth of detail about her familial, political, professional, religious, and sexual associations”, as noted by Justice Sotomayor. [2] This collection of data was determined to have violated the fourth amendment. The argument of this being an organised crime may be valid as well, which would lead to further legal complications. Even though this would come under a cyber-orientated risk, there could be a risk of actual physical harm. The use of precise information could be exploited and revealed to malicious parties, which could potentially result in stalking and further violence. Thus, depending on the situation, autonomous vehicles could be assisting such criminal offences. An equally concerning and alarming issue is that this could also lead to physical harm in instances related to hackers and cybercrime, where hackers could manipulate self-driving cars, exploiting vulnerabilities in LIDAR technology which could lead to collisions or other criminal offences.

Additionally, there is also the concern about the collection of sensor data, which means that autonomous vehicles would have records of environments and have the use of cameras, sensors etc. Even though many companies already collect and use sensor data, such as Google, where the FCC sanctioned the transmission of sensitive data through Wi-Fi in 2012, the future collection could result in an invasion of individual privacy as such data could lead to autonomous vehicles predicting daily habits and destinations and may even reveal such information about the user without proper consent. [3] In order for the protection of the privacy rights of an individual, policies and safeguards must be introduced, which would ensure the protection of the data that would be collected by autonomous vehicles.

The Collingridge dilemma could be used as an example when it comes to cybercrime and data protection. New technology could bring a ‘dilemma of control’, where technology would be innovative and advanced, however, this could lead to less control over it and more cybercrime. The efforts to create more advanced forms of technology could result in its downfall. If there is a major development in autonomous vehicles too quickly, this could receive negative attention and develop to an advanced extent without proper regulations in place, making AV systems more of a hazard and an economic issue. For an emerging technology such as AVs, it is important to avoid the Collingridge dilemma, which could result in large amounts of cybercrime. [4] However, on the contrary, there is the issue that if policymakers continue to hold back the emergence of AV systems, this would result in suppressing the development in general. The question arises of when is too early and when is too late for policymakers to intervene and speed up the process.

Unemployment

An ethical challenge that would be faced while assessing autonomous vehicles is the issue of rising unemployment. Contemporary society is seeing an uprising of technology replacing jobs, such as self-service machines and robotics in factories, which results in low hiring rates and high unemployment. This would also be evident in autonomous vehicles, as much employment and income is of those who drive vehicles, for example, deliveries, taxi and truck drivers. These jobs account for 370,499 jobs for deliveries and taxis in America, there are more than 680,000 bus drivers who are employed, which reflects the possible scope of possible loss of jobs in just America alone. [5] They all rely on their income through driving vehicles. Autonomous vehicles would result in the income of future generations being restricted, which is where the ethical challenges come into place. Despite there being a small number of solutions to this upcoming issue, there could be an alternative way to provide a source of income in the use of these vehicles. For example, the autonomous vehicle industry itself would actually be creating jobs and hiring more individuals, this would be in many aspects of many different fields, from software to manual work. This has been evident, as in January 2018, there was a 27% increase in job listings referenced to autonomous vehicles, which was a dramatic increase from 2017. [6] So although the ethical argument of unemployment arises, there are also arguments for higher employment being raised in parallel.

Licensing 

The framework based around autonomous vehicles is that they are self-driving vehicles, however, this brings up the question of whether acquiring a license would be necessary for the user of such a vehicle. There is also the question of whether users with physical and mental conditions should be allowed to use these vehicles if they would not originally be allowed a license. Moreover, there is also the argument that further training may actually be necessary, as autonomous vehicles operate differently from a standard car, for example, training in safety and understanding new safety features. The legal framework based around vehicle licenses would need to be reformed and new legislation, whether this be a need for a license with an autonomous vehicle or not, would have to be added in order to update procedures. Currently, when owning a Tesla in the UK there is the requirement for a license because they do not operate as fully autonomous, however in the US, even though they are mostly autonomous vehicles, a license is still required. At least for the next decade or so, because AV systems have shown to fail highly in autonomous systems.

Liability 

A major issue in relation to autonomous vehicles is the one based on liability. This is both a legal and ethical issue, however, the main focus would be the stance of ethics. Autonomous vehicles are still developing and even though AV systems decrease the instances of car crashes, this does not mean that there aren’t any, which is where the question of liability pivots. When a car crash occurs who should be held liable, especially if there is a life that was lost? The driver would be less liable as they are not driving the car, however, technology would not be blamed, so it could be the manufacturers. It has been shown that manufacturers are generally more conservative when programming autonomous vehicles to avoid collisions and thus avoid liability. When questioned with liability, the legal and ethical issue of criminal liability emerges. There is the argument that the user who is in charge of the AV would not be held accountable if the vehicle was fully autonomous during the collision. Legislation based around liability for the user would be difficult, especially under ‘act of omission’. However, legislation could be based upon the use of reliance and there could an assessment of whether the user had the opportunity to take control of the situation. In the Law Commission report published 2022, on automated vehicles, there is the argument that criminal liability should be held for senior managers. [7] Many consultees stated that senior staff should be held liable, although there is an argument that they may hold the highest technical knowledge, but they may not have a direct responsibility in relation to safety. The conclusion that the law commission came to was that senior staff could face prosecution if the offence was under their ‘consent and connivance’. This exhibits an extensive view into the legal framework of liability in connection to autonomous vehicles, as it is evident that the UK legal framework has come to a partial conclusion to liability and a partial fix for the trolley problem. However, it is important to mention that, on the contrary, if an individual was to interfere with an autonomous vehicle which would result in injury and serious harm to property, this would also be considered a criminal offence and the user would be held liable, as they interfered. This is stated in the Law commission automated vehicles joint report, wherein Scotland, Section 22A states that if an individual interferes with an AV which results in injury or even serious damage to property it would be considered an offence. This shows the upcoming legislation which looks into both sides of autonomous vehicles and liability. This section used MacPhail V Clark 1983 as an example of the flexibility of such an offence. [8]

However, even though there is the argument that humans must stay alert to intervene in emergency instances, where a collision may occur, it is important to note that humans take time to come to situational awareness, especially if they have not been focused on the road. The awareness aspect of autonomous vehicles also defeats the purpose of their creation as level 4 and level 5 AV systems. Placing liability on the AV users would allow for the manufacturers to get away with misleading marketing and lead to them being able to displace AV liability, as the reasoning behind an AV user would buy an autonomous vehicle would be to be able to have a safer and self-driving journey home, and if they take control of the vehicle while the collision was in process, they would then be held liable. Thus, the AV user should hold less liability than the manufacturer.

Even though there could be a way to assemble legislation based around the user where there is criminal liability, there would also be the need to scrutinise the manufacturer and investigate their criminal liability, probing into more legislation based around this. Another important aspect of regulating the law in relation to autonomous vehicles is analysing how the victims would be compensated. This would be an area of further legislation as traditional lawsuits would not be valid as many laws are vague in the AV sphere and compensation would also be vague, therefore there would be a need for alternative victim compensation.

However, liability does create a new spectrum in the insurance sphere and many issues. This also may interconnect with the argument of liability. If the user of the autonomous vehicle was in control of the vehicle during the collision this would allow the insurance company to provide a personal auto-coverage, however, the question arises if the autonomous vehicle has full control. However, autonomous vehicles could prove to be more beneficial for insurers as around 94% of car crashes occur purely because of human error and the use of autonomous vehicles means the likelihood of accidents decreases, which is financially beneficial to insurance companies. [9] There is the issue which would occur of drinking and driving, as research from Curtin University, Perth, Australia conducted a survey showing that the rise of autonomous vehicles would increase binge drinking. [10] This creates a new set of challenges for the legislation based on autonomous vehicles, because lawmakers would have to take into account the risks of these vehicles which would bring in negative consequences. Even though there are ways that lawmakers could try to prevent this which would be keeping drunk driving even in an AV illegal, it would still result in an introduction to even more negative health effects.

The role of law in relation to the regulation of liability of autonomous vehicles internationally is very little. Currently, governments like China and South Korea have decided to not express any response on their legal framework of the regulation of autonomous vehicles, although, countries like Singapore have decided to start working on regulation of autonomous vehicles and updated liability legislation, they decided to amend a RTA in 2017 which would exempt autonomous vehicles involved in the AV trials which were already in place by the RTA, in which it was stated that the user would be held accountable for the use of the vehicle on public roads. [11] This shows that there is a shift of liability, and that Singapore acknowledges that the user would not be in control of the AV system. The European Commission in 2016 had expressed that they were looking into autonomous vehicles legal framework as they had launched GEAR 20130, which had expressed this. Although the UK has shown to be pro-active when it comes to legalisation and autonomous vehicles, they published the law commission report on automated vehicles and had particular legislation in place, such as the consumer protection Act which protected manufacturers, if they had shown enough evidence that their products did not have any defects.

 Furthermore, insurers could use data from autonomous vehicles to calculate liability and collisions, as autonomous vehicles are largely reliant on data recorded in high volumes and are usually compliant with data protection laws.  This does create a legal issue, as insurers would have a right to the data and this could violate data privacy for the users of the vehicle. A 2016 report, conducted by the Platform for the Deployment of Cooperative Intelligent Transport Systems, concluded that vehicle identification could result in access to personal data, which causes privacy concerns and a need for a legal framework. [12]

Regulations

When analysing ethics and laws in association with autonomous vehicles, assessing past and current regulations is a necessary part. There have been international legal instruments that have addressed and looked into the regulations of autonomous vehicles. However, it is also important to relate older legislation. In 1949 at the Geneva Convention on Road Traffic, the members had come to the mutual agreement of furthering road safety internationally. In article 8 it is stated that ‘every vehicle shall have a driver’, which is important to acknowledge when looking into regulation based around autonomous vehicles, as autonomous vehicles have the ultimate goal of being self-driving, which could create a legal issue with manufacturers when they must meet this legal standard internationally, as it could affect them financially and defeat the ideology of their technology. [13] The international legal concept is challenged when the autonomy of the AV reached a higher tier. Currently, autonomous vehicles would be placed in a level 2, which is partial automation, this can be seen in cars such as Tesla. Although this does not mean that international legal parties are completely dismissing this legal issue, at the 1968 Vienna Convention on Road Traffic, amendments were proposed and have been since looked into. These would be based upon domestic technical regulations and domestic legislation governing operations, which opens up the roads for autonomous vehicles when introduced and allows further legislation for autonomous vehicles if the regulations are met.

Moreover, in terms of regulation, there is the legal framework of the market in relation to autonomous vehicles to assess, where the safety of AVs would have to be authorised before being able to be established onto the market. The regulations with assessments and safety in use would be an ongoing issue, therefore the legal framework would have to be thorough and express validity.  There is also the need for marketing regulation as many autonomous vehicles may be marketed as self-driving whereas they cannot, they may be a level 2 AV marketed as a level 4 when they are not authorised under the correct government scheme, and this could lead to misleading marketing and road-safety issues, as many users may fail to recognise the car’s abilities and then reduce their attention to the road. This legal issue has been brought to the attention of the law commission and was looked into in the ‘Automated vehicles: joint report’, where they mentioned that they recommended two new criminal offences which would help restrict the use of the label ‘self-driving’ in the market. [14] This illustrates that the UK is working on a legal framework for developing autonomous vehicles and introducing them while considering many aspects when it comes to regulation.

The trolley problem

When looking into the regulation of autonomous vehicles a significant aspect would be the ethical conundrum of the trolley problem. As technology such as autonomous vehicles can also have errors and run into unpreventable environmental problems, there are issues that must be faced where the vehicle makes a choice. The trolley problem is a psychological problem that questions morals and the justification of murder in certain situations, for example, would it be valid to kill 1 life to save 4? There is a large debate, purely because this is such an ethical aspect of autonomous vehicles, which means that there is arguably no correct response to the trolley problem, but there are some arguments for and against it. The reason the trolley problem is crucial in the ethical regulation of autonomous vehicles is that the question arises of what a self-driving vehicle should do if they are given the choice in a situation to save different lives. Should the AV make the choice of crashing into a light post and killing the passenger or should they continue and crash into people crossing, resulting in their deaths. How should the autonomous vehicle be programmed to make the correct decision when the decision is an ethical debate? Additionally, there is the role of anti-discrimination law such as the anti-discrimination act 2010, where it is important to note if policymakers would actually take the necessary step to look into the software of the autonomous vehicles and prevent any form of discrimination. Racial bias within AI and other developing technologies has been evident, this is because they rely on constant data, which would be interpreted by the software. An example would be whether an autonomous vehicle which mostly is exposed to white individuals in a majority white neighbourhood, would choose to save the life of a white individual over the life of a person of colour, as it would have collected data that would create a bias. [15] Specific characteristics, such as gender, age and even size would fall into an AV algorithm which could cause other groups to be given privileges and this would create a form of discrimination.

There is also the argument that the manufacturers could use these algorithms and intentionally protect particular classes which could benefit them. The very decision making that the manufacturers have to embed within the autonomous vehicle would be largely based on ethical preferences unless legislation introduced states otherwise. The manufacturer’s ethical preferences have already been evident in the past, as seen with Mercedes and their developing autonomous vehicles, they based their statement on level 4 and 5 autonomous cars and said that they would sacrifice the pedestrians and save the passenger’s life in every situation. [16] However, even though this is what the manufacturers have said, it would be what the law permits, which is where the question of whether the law would even permit autonomous vehicles to be programmed with ethical and moral rules. AI and any form of robotic technology would mostly not follow moral conduct on its own but instead follow one which is more beneficial economically by decreasing liability. Although, under negligence, there is the role of law that legislation could permit that AV could be programmed to collide with them, as they hold liability as well.

Furthermore, autonomous manufacturers would base their programming on particular incentives as we, for example, manufacturers understand that stakeholders would hold different incentives which would motivate them to invest, which means the very framework that the AV algorithm is programmed into would maximise profits. This would mean that the trolley problem would not be an issue that would be faced, instead the use of economic incentive v morals would be placed. These biased incentives could create major implications in the legal framework of autonomous vehicles.

To conclude, the role of law in connection to autonomous vehicles is very minimal in general internationally, although, the UK legal framework has looked into and expressed more insight into AV systems and therefore created a basic framework, they have shown an interest in developing this further, as expressed in the law commissions joint: report on automated vehicles. There are many legal issues, such as licensing, insurance, and liability where amendments and new legislation are necessary in order to introduce level 4 and 5 AV systems. Further issues would be based on policymakers and the question of how quickly they should introduce regulation as this would correlate to the introduction of advanced AV systems. The main ethical conundrum when looking into AV systems is the trolley problem, which when researched showed that the outcome could be manipulated, by different parties to create the outcome best for them instead of what would morally be the ideal solution when programming. The issue of the level of safety of AV systems before they can be introduced on the road, which was concluded to be 2 or 4, as level 3 would not be possible due to environmental and emergency hazards. The law and ethics on AV systems are gradually becoming a vastly researched topic for governments, which opens more insight into legislation.

Bibliography

1. Gerber Gerber Law, 'Autonomous (Self Driving) Vehicle Statistics 2021' (Gerber Injury Law, 2022) <https://www.gerberinjurylaw.com/blog/2021/june/autonomous-vehicle-statistics/> accessed 2 March 2022.

2.   'United States V. Jones | Case Brief For Law School | Lexisnexis' (Community, 2022) <https://www.lexisnexis.com/community/casebrief/p/casebrief-united-states-v-jones> accessed 6 March 2022.

3.   Viano E, Cybercrime, Organized Crime, And Societal Responses (Springer 2016).

4.   'The Privacy Implications Of Autonomous Vehicles' (Data Protection Report, 2022) <https://www.dataprotectionreport.com/2017/07/the-privacy-implications-of-autonomous-vehicles/> accessed 16 February 2022.

 5.  C.C & J.S, 'Reframing The Governance Of Automotive Automation: Insights From UK Stakeholder Workshops' (Taylor & Francis, 2018) <https://www.tandfonline.com/doi/full/10.1080/23299460.2018.1495030> accessed 4 March 2022.

 6. U.S. Bureau of Labor Statistics (BLS). "Occupational Outlook Handbook: Heavy and Tractor-Trailer Truck Drivers." Accessed May 18, 2021.

 7.  (2022) <https://www.cnbc.com/2018/08/10/autonomous-vehicles-are-creating-jobs-heres-where.html> accessed 1 March 2022.

8.   Law Commission, 'Automated Vehicles : Joint Report' (S3-eu-west-2.amazonaws.com, 2022) <https://s3-eu-west-2.amazonaws.com/lawcom-prod-storage-11jsxou24uy7q/uploads/2022/01/Automated-vehicles-joint-report-cvr-03-02-22.pdf> accessed 9 March 2022.

9.   L. Song Richardson, 'WHEN HUMAN EXPERIMENTATION IS CRIMINAL' (Jstor.org) <https://www.jstor.org/stable/20685033?Search=yes&resultItemClick=true&searchText=macphail%20v%20clark&searchUri=%2Faction%2FdoBasicSearch%3FQuery%3Dmacphail%2Bv%2Bclark%26so%3Drel%26efqs%3DeyJjdHkiOltdLCJkaXNjIjpbImJHRjNMV1JwYzJOcGNHeHBibVU9Il19&ab_segments=0%2Fbasic_search_gsv2%2Fcontrol&refreqid=fastly-default%3Aaaed980430972195c14e8050dc95f810> accessed 1 March 2022.

10. Chris J Roy Jr (Hg.org) <https://www.hg.org/legal-articles/car-accidents-are-almost-always-due-to-human-error-50458> accessed 3 March 2022.

11.   (Autoexpress.co.uk, 2022) <https://www.autoexpress.co.uk/car-news/105884/self-driving-cars-will-lead-to-parking-chaos#:~:text=Driverless%20cars%20will%20reduce%20drink,driving%20age%20who%20consume%20alcohol.> accessed 1 March 2022.

12.   Taeihagh A & Lim H , 'Governing Autonomous Vehicles: Emerging Responses For Safety, Liability, Privacy, Cybersecurity, And Industry Risks' (Taylor & Francis, 2018) <https://www.tandfonline.com/doi/full/10.1080/01441647.2018.1494640> accessed 18 February 2022.

13.   (Pinsentmasons.com, 2022) <https://www.pinsentmasons.com/out-law/analysis/car-insurance-evolve-autonomous-vehicles> accessed 3 March 2022.

 14.  United Affairs, 'No. 1671. Convention On Road Traffic. Signed At Geneva, On 19 September 1949 | United Nations Ilibrary' (Un-ilibrary.org, 1998) <https://www.un-ilibrary.org/content/books/9789210596640s003-c003> accessed 2 March 2022.

15.  Law Commission, 'Automated Vehicles: Joint Report' (S3-eu-west-2.amazonaws.com, 2022) <https://s3-eu-west-2.amazonaws.com/lawcom-prod-storage-11jsxou24uy7q/uploads/2022/01/Automated-vehicles-joint-report-cvr-03-02-22.pdf> accessed 9 March 2022.

  16. Taeihagh A & Lim H, 'An Examination Of Discrimination And Safety And Liability Risks Stemming From Algorithmic Decision-Making In Avs' (Ippapublicpolicy.org, 2019) <https://www.ippapublicpolicy.org/file/paper/5cec300e61a91.pdf> accessed 5 March 2022.

  17. 'Self-Driving Mercedes Will Be Programmed To Sacrifice Pedestrians To Save The Driver' (Fast Company, 2022) <https://www.fastcompany.com/3064539/self-driving-mercedes-will-be-programmed-to-sacrifice-pedestrians-to-save-the-driver> accessed 9 March 2022.

Cases

MacPhail V Clark 1983 SCCR 395

United States v. Jones, 565 U.S. 400 (2012)

Legislation

Scotland Section 22A

S1 Automated and Electric Vehicles Act (2018)

1949 at the Geneva Convention on Road Traffic, article 8

US constitution, Fourth Amendment

Vienna Convention (1969)

Law Commission joint report (2020 publication)

Downloadable version :

Legalities of Autonomous Vehicles-JiyaBharti.pdf

Download PDF • 194KB

Thank you for reading! I hope you enjoyed it!