Keywords

1 Introduction

The impact of AV on governance is not well researched yet. The aim of this chapter is to raise first insights on how governance should be adapted to the upcoming change in the mobility paradigm based on AVs as a game changer. In this chapter, the impact of automated vehicles (AVs) on legislation will be briefly exposed, as well as the types of governance that are and could be used for their actual and future regulation. A second part will briefly present a selection of legal recommendations to public authorities that were drafted within the AVENUE project.

1.1 The Deployment of AVs Will Impact All Areas of the Law

Automated vehicles (AVs) and the mobility services they provide will require changes in all areas of the law:

  • Administrative/public law (through modifications to road traffic rules and regulations (World Economic Forum, 2019, p. 5))

  • Civil law (amongst others by updating the regulations of civil liability and product liability in cases of accidents (PUNEV, 2020, p. 99))

  • Criminal law (for the case of accidents with bodily injuries, misuse cases of AVs and, potentially, for infringements of road traffic rules committed by AVs (Council of Europe CDPC, 2020))

These areas of the law can be either:

  1. 1.

    Strongly harmonised and regulated at international and/or regional (European) levels, with small to no room for different material rules (e.g. having rules with different content) at national level

  2. 2.

    Mostly or completely regulated by national legislation, which means the material legal rules are defined at a national level and can therefore be different from country to country (for instance, the position of the driver’s wheel on the left or ride side of the vehicle)

  3. 3.

    A mixture between both, with for instance international principles leaving room for national differences (UNECE, 2022)

1.2 Types of Governance

As the deployment of AVs has the potential to significantly impact all areas of the law, questions related to the best or more efficient governance and regulatory frameworks arise.

Governance can be described as the process of making and enforcing decisions within an organisation, state or society. It is the process of interactions through the laws, social norms, power (social and political) or language as structured in communication of an organised society over a social system (family, social group, formal or informal organisation, a territory under a jurisdiction or across territories). It is done by the government or the parliament of a State, by a market or by a network. It is the process of choosing the right course amongst the actors involved in a collective problem that leads to the creation, reinforcement or reproduction of acceptable conduct and social order (Möltgen-Sicking & Winter, 2019).

As regards the word “law”, we will use the following terminology that makes the distinction between:

  • Laws in the formal sense, i.e. norms that are adopted by the State’s legislative body (e.g. the parliament) in the course of the regular legislative process (entitled “law”).

  • Laws in the material sense, i.e. general and abstract norms that can be adopted by other bodies than the legislative, such as the State’s executive body (e.g. the government). They can be called regulations, ordinances, acts, rules and so on.

Here are some example of different approaches to governance and regulation (International Association of Public Transport, 2021):

  • Global governance: Global governance involves the creation of international agreements, treaties and institutions to regulate the development, deployment and use of new technologies on a global scale, providing harmonisation such as seen above. This type of governance is particularly important for technologies that have cross-border implications, such as AVs. In the field of automotive, the UNECE and other bodies are already playing this role and are working on topics related to AVs (Lagrange, 2021).

  • Governmental regulation: This type of governance involves the creation and enforcement of laws and regulations by government bodies to control the development, deployment and use of new technologies. Governmental regulation is seen as necessary to protect public safety and security, privacy and other public interests. In the case of automated vehicles, governmental regulation will be necessary for areas such as road traffic safety, cybersecurity, data protection and liability, for instance (World Economic Forum, 2019, pp. 8 et seq.). Governmental regulation at state level takes place into global governance, where it can be partly defined by binding international instruments. In the field of new technologies, the European Commission has already enacted various acts regulating amongst others the processing of data (such as the General Data Protection Regulation, the upcoming Data Act, the Data Governance Act, the Digital Markets Act, the Digital Service Act, the AI Act etc.).

  • Self-governance: This type of governance is based on the principle of voluntary compliance by individuals or organisations that are involved in the development, deployment or use of new technologies such as AVs. Self-governance may involve the creation of industry standards, codes of conduct or best practices that guide the behaviour of the stakeholders in a particular industry. For AVs, the interests of public safety and security (amongst others) do not allow relying on self-governance alone, and governmental regulations are required. In particular fields, where safety and security are not at stake, lawmakers and governments might allow the stakeholders (and the markets) to regulate themselves, following a so-called laissez-faire strategy.

  • Collaborative governance or co-regulation: This involves collaboration between lawmakers (e.g. governments and legislative bodies), industry, researchers and other stakeholders to develop and implement policies and standards (Law Commission of England and Wales and the Scottish Law Commission, 2022a). This approach should be followed for AVs, especially at this early stage of AV deployment, as it helps ensuring that the interests of all stakeholders are taken into account and can promote greater acceptance of automated vehicles amongst the public. Safety, with, for instance, the definition of passive safety requirements (and the necessary trade-offs related to vehicle architecture), might be solved efficiently through collaborative governance (International Association of Public Transport, 2021).

  • Social governance: Social governance refers to the use of non-governmental organisations, community groups and other civil society actors to shape the development and deployment of new technologies. Social governance may involve public education, advocacy or activism to promote responsible innovation and ensure that the benefits of new technologies are shared equitably. Social governance will be needed to ensure the acceptation of AVs by the public and to ensure, for instance, that PRMs get involved in the development and will benefit from AVs.

  • Technological governance: Technological governance involves the development of technical standards and protocols that ensure the interoperability and compatibility of automated vehicles with other systems and technologies. This approach can help ensure that automated vehicles can operate safely and efficiently within existing transportation systems. In the field of AVs, AVENUE, for example, has shown the necessity to provide open datasets and application protocol interfaces (APIs) for the common good.

  • Financial governance: Financial governance involves the development of policies and regulations related amongst others to the financing of automated vehicles, such as tax incentives or subsidies for the development and purchase of automated vehicles by public transport operators. Regulation also plays an incentive role, as “strict” regulations with high type-approval costs or triggering high costs for the deployment of automated vehicles have an important financial impact for vehicles manufacturers or public transport operators. Besides, it should be noted that an optimal use of AVs might require massive investments in public infrastructure.

During AVENUE, voices have been raising the fact that European regulations, for instance, in the field of type-approval and/or getting an approval for AV experimentations, as well as in the field of data privacy with the General Data Protection Regulation, tend to be more restrictive than other regulations (viz. in the USA or in Asian countries). According to these, from an economical point of view, this would result in a comparative disadvantage for research and vehicle manufacturers in Europe. A counter-argument is that, in the long run, these higher standards will also support achieving better products, output and results, increasing safety and security and avoiding unnecessary incidents. As of today, according to the author’s opinion, the approach to vehicle automation followed in the USA has resulted in more accidents with body injuries or even death cases involving advanced driver assistance systems or automated driving systems than in Europe (Tesladeaths, n.d.).

Overall, effective governance of automated vehicles will likely require a combination of these different approaches, tailored to the specific needs and characteristics of different regions and countries but also to the specific AV domain to regulate.

1.3 Major Impacts on Legislation and on the Societal Activities they Regulate

The deployment of automated vehicles will have major impacts in many fields involving public authorities and interests, far beyond formal regulations provided by public, civil or criminal laws:

  • Legal and regulatory framework: As mentioned before, the introduction of automated vehicles will require a re-evaluation of many existing legal and regulatory frameworks, as it will be briefly summarised in the following chapters. Lawmakers and executive bodies at different levels will need to consider new international and national regulations and instruments to address the unique characteristics of automated vehicles. AVs are part of the upcoming technological revolution that will be triggered by artificial intelligence and the internet of things which will require updates and changes in many existing regulations as well as the creation of new legal frameworks.

  • Infrastructure and urban planning: Automated vehicles will require new infrastructure and urban planning (Martin et al., 2023) to support their safe and efficient operation. They could also completely change the landscape of our roads and cities. Some changes have already begun: vehicle manufacturers are developing the use case of “Remote Controlled Parking” systems (RCP), where an “Automatically Commanded Steering Function” (ACSF) of the vehicle performs the dynamic driving task (DDT) of the vehicle without any human intervention. In this use case, the driver and passengers get out of the vehicle, which is empty of any human driver and parks itself into a free parking lot “alone” by using the ACSF. This allows gaining parking space, as vehicles equipped with RCP can park themselves in much narrower spaces than the existing ones, where space is needed for the driver and the passengers to open the doors of the vehicle in order to enter or exit the vehicle (Huonder & Raemy, 2016).

  • Public investments in new technologies and infrastructure will be needed, forcing lawmakers and executive bodies to define the AV environment they wish for their roads, such as the type of communication networks (vehicle-to-vehicle, vehicle-to-infrastructure etc.) and related sensors and communication channels (Nationale Plattform Zukunft der Mobilität, 2020), charging stations etc. to support the deployment of AVs. These developments will also be pushed by the fact that conventional vehicles are becoming connected vehicles, and AVs are a subcategory of connected vehicles. In this field, decision-makers will have to proceed to technological, political and financial trade-offs, for instance, whether they want AVs that do not rely on sensors placed in the road infrastructure or whether the road infrastructure should be able to communicate with AVs (National Plattform Zukunft der Mobilität, 2021a, 2021b) to provide them information, and/or whether redundancies are needed, in particular for the sake of safety and security. An example: using its sensors, a fully automated vehicle could detect on its own whether a traffic light is red or green, but an “intelligent” or connected traffic light could also communicate through wireless communication channels if it is red or green and provide this information to the AV. These decisions will have important financial impacts on public budgets and might also impact the rhythm of AV deployment, by accelerating or slowing it down.

  • Employment and workforce: The introduction of automated vehicles will have a significant impact on the transportation industry, leading to inevitable job losses in certain sectors such as public transportation, trucking and taxi services (International Association of Public Transport, 2017). Governments will need to consider the impact on these workers and develop policies to address the potential disruption. Driver unions have, for instance, started to protest against the development of automated vehicles in California in the beginning of 2024 (TheLastDriverLicenseHolder, 2024a), and a few cases of attacks against automated vehicles have been reported recently in San Francisco (TheLastDriverLicenseHolder, 2024b). In parallel, AVs will also create new jobs that didn’t exist previously.

  • Social and economic impact: Automated vehicles will significantly impact society and the economy, by (possibly) reducing traffic congestion, improving safety and increasing mobility for persons with reduced mobility. Governments will need to consider the potential social and economic impact of automated vehicles and develop policies to support their deployment in a way that benefits society as a whole (Smith, 2016). According to the research of the AVENUE projects, some current forms of AVs, such as robotaxis or private automated vehicles, will not benefit to society as much as shared and automated forms of public transportation, which should be sustained by public authorities.

  • Ethical considerations: The development and deployment of automated vehicles also raises ethical considerations, such as how to program the automated driving system to take decisions in situations where competing interests are at hand, as described in the famous “trolley dilemma” (Foot, 1978) where an automated driving system would need to choose between hurting the AVs’ passengers or hurting third parties such as pedestrians on the road. Governments will need to consider these ethical considerations and develop policies to ensure that automated vehicles operate in an ethical and socially responsible manner. Germany, for instance, has already tackled this question with its Ethics Commission that released guidelines for AVs in 2018 (Ethik-Kommission, 2017).

2 AVENUE Recommendations

The following sections will briefly present the most important legal recommendations for public authorities that were drafted during the AVENUE project. They enter into the categories of global governance, governmental regulation and collaborative governance or co-regulation defined above.

The first section will be dedicated to cross-sectoral recommendations, followed by another dedicated to specific topics.

2.1 General Recommendations Not Limited to Specific Areas of the Law

2.1.1 Prioritising the General Interest Through Encouraging Legislation

Roughly summarised and without going into the details of the SAE Levels, the current development of AVs can be categorised in three different approaches or use cases:

  • Privately owned automated vehicles, e.g. conventional vehicles with a driver and a steering wheel, improved with advanced driver-assistance systems (ADAS) or automated driving systems (ADS), answering private mobility needs of individuals owning these vehicles.

  • Robotaxis, which are automated vehicles that can be similar to privately owned vehicles, but are owned and deployed by commercial companies with the purpose of providing ride-hailing services and generating profits for the companies.

  • Shared automated vehicles integrated in a MaaS (Mobility as a Service) or an ITS (Intelligent Transport System), often falling into the category of minibuses/shuttles, deployed by transport operators providing traditional transport services and/or used door to door and on demand in a MaaS or an ITS as mobility gap filler to complement private or public transport services enabling positive externalities (see Chap. 18 about vision).

If we look at the current regulations being adopted by public authorities in the field of automated vehicles, the current lawmaking momentum lies mostly with the regulation of the first two categories, namely, privately owned automated vehicles and/or robotaxis:

  • The first AV-specific regulation enacted at international level (as a result of global governance defined above) was the regulation dedicated to the Automated Lane Keeping System (ALKS), a (SAE Level 3) highway traffic jam assistant designed for conventional, privately owned vehicles with automated driving systems. It was enacted by the United Nations Economic Commission for Europe (UNECE) in 2020 and consolidated to 130 km/h in 2022 (UNECE, 2022). Since then, in a few European countries, ALKS has been type-approved and can be lawfully used on certain highways (KBA, n.d.).

  • At national levels, most of the experimentations and legislative output are focused on the regulation of robotaxis (see, for instance, in San Francisco with Cruise and Waymo amongst others (TheLastDriverLicenseHolder, n.d.)).

This can be explained by the fact that these types of vehicles (e.g. privately owned vehicles with ADAS or ADS or robotaxis) and their use cases are promoted by private stakeholders with very important resources wishing to enter in what they consider a new promising market. These stakeholders are either “tech” companies (such as Waymo, a company curated by Alphabet/Google (Waymo, n.d.), or Mobileye, curated by Intel (Mobileye, n.d.)) or manufacturers of conventional vehicles vying to join the automation trend (such as Mercedes (The Verge, n.d.), Cruise/General Motors (Cruise, n.d.) etc.).

They can also be both type of actors, working together in the form of joint ventures and consortiums. As a consequence, the legislation enacted by legislative bodies and governments follows the technical developments proposed by these stakeholders and reacts to it with a “bottom-up” approach, similar to collaborative governance, although often the public authorities lack of the necessary knowledge and involvement.

Shared automated minibuses, which are, according to the AVENUE research results, the more desirable automated driving use cases from a societal point of view, did not stand in the focus of lawmakers until recently.

Thus, priority should be given to their regulation, as research within AVENUE (and other projects as well (Fabulos, n.d.)) has demonstrated that shared automated minibuses are the use case of AVs that will provide greater and better societal benefits (International Association of Public Transport, 2021) than privately owned automated vehicles or robotaxis, which might trigger negative externalities such as an increase in traffic congestion and weaken public transportation and accordingly generate additional costs for public transport authorities.

Therefore, governments and lawmakers should play an orientating role by encouraging the most promising use cases and by creating incentives. Accordingly, they should also regulate and mitigate use cases considered to be less desirable from a societal point of view.

2.1.2 Cooperation and Harmonisation at European Level

For the development of automated vehicles, at European level, an aligned and holistic legal European approach towards automated vehicles is desirable (European Commission, 2020a).

As seen during the AVENUE project, the same automated minibuses went through very different approval proceedings in the four countries of experimentations, with important time differences, to receive an approval from public authorities. The cost and documentation required from each of the partners in France, Luxembourg, Denmark and Switzerland were also very different, although the automated minibuses were of the same model and manufacturer.

Therefore, an aligned and holistic legal European approach is desirable and necessary for many reasons, starting from road safety to a commercial, single market perspective, with the promotion of the European industry and encouragement of technological leadership. Such an approach and harmonisation can be defined through European institutions and regulations, which have yet to be created and defined.

2.1.3 Cooperation Between All Stakeholders

In order to cope with the fast evolving technologies and business models involving automated vehicles and minibuses, various governance models can be implemented, such as collaborative governance or co-regulation, where all stakeholders get involved in the definition of the regulation/policy through an inclusive approach and with the sharing of knowledge and resources (International Association of Public Transport, 2021).

Such approaches are needed as private stakeholders (vehicle manufacturers, “tech” companies, software developers etc.), public transportation operators (PTOs), public transportation authorities (PTAs) and further stakeholders (organisations in charge of accident analysis, insurers etc.) already involved in the deployment of AVs can identify their needs and legislative issues through their experience already gathered while experimenting.

Similar results might also be achieved through self-governance of non-institutional platforms and/or private associations, representing the interests of the various stakeholders. Harmonisation through public institutions or involving public institutions might be more desirable, as private associations might follow and defend the interests of their members that may differ and strive for different goals such as maximisation of profits. This might result in unsatisfying results from a societal point of view, as the potential positive societal benefits and outcomes studied in the AVENUE project might not be reached.

2.1.4 Coordination Between Authorities

AVs will stand in the focus of different governmental authorities (Law Commission of England and Wales and the Scottish Law Commission, 2022b; Monti & De Streel, 2022), as they function based on multiple technologies supervised by different authorities: type approval authorities for the “road vehicle” components, telecommunication authorities for the use of antennas and radiocommunications, data protection authorities as regards the processing of personal data and even competition authorities in the case of potential concentration of market power by actors involved in the AV chain.

As of today, these authorities rarely work together. These authorities will also need to coordinate themselves and develop new skills and competences related to digitalisation and automation.

2.1.5 Standardisation

As a result of self-governance, the vehicle industry often relies on proprietary standards and formats (for instance, each vehicle manufacturer uses its own format of a high definition map, which is generally not compatible with other manufacturers’ maps). This creates high curation costs when data is transferred from one company to another and high costs to develop new services (Iacob et al., 2021; Easme & Cosme, 2018). Standardisation and interoperability in all domains should be encouraged as much as possible.

2.1.6 Definitions and Vocabulary

The vocabulary used to describe automated driving technologies should not mislead the public and create false expectations. A certain level of harmonisation that would increase the public’s awareness and comprehension of these new technologies is desirable (Nationale Platform Zukunft der Mobilität, 2021b).

2.2 Recommendations Related to Road Traffic Rules and Conventions

2.2.1 Amendments to the Existing Road Traffic Rules and Conventions

During many decades, the presence of a human driver, located inside the vehicle and performing the “dynamic driving task”, was a central assumption of the international road traffic conventions such as the Vienna and Geneva Conventions, as well as for many national road traffic laws. With the deployment of AVs, a fundamental paradigm change is currently operating, and states should depart from this twentieth-century approach to driving (Smith, 2022; Vellinga, 2019).

At international level, the UNECE is currently working to deliver a regulation for ADs in 2026 (Pichereau, 2021). However, the timeframe for the final adoption for such regulation might be too slow and take too long before the effective deployment on AVs on the roads.

Each country should therefore proceed as soon as possible to an extensive assessment of its road traffic regulations, in order to prepare and amend:

  • Legal or technical provisions that implicitly or explicitly relate to the presence of a human driver performing the dynamic driving task from inside the vehicle

  • Legal or technical provisions that might pose interpretation problems with the introduction of automated vehicles

  • Legal or technical provisions involving human negotiation, interpretation or communication in specific situations (for instance, when a policeman is regulating the traffic at a specific spot through hand signals)

It should also be noted that automated vehicles are programmed to comply strictly with the existing traffic rules and not infringe them. In the road traffic reality with human drivers, however, various situations can occur that require a trade-off between obstructing the traffic flow and infringing road traffic rules. For instance, after an incident with the damaged vehicles located on the lane, vehicles driven by human drivers will carefully cross the lane to circumvent the obstacle. This simple arbitration might be difficult to implement in the case of automated vehicles, as this solution might imply to infringe traffic rules.

Solutions to such problematic situations requiring trade-offs and arbitrations should be discussed with the various concerned authorities (such as road safety, police, traffic law, politicians etc.), the developers of automated vehicles and researchers.

2.3 Supervision of Automated Vehicles

2.3.1 Allow Remote Supervision of AVs

Having a safety operator located inside each automated minibus (as required by today’s experimentations’ approvals) generates important functioning costs and does not constitute an economically viable solution in the long term for commercial and large-scale deployment (Probst, 2021). Therefore, legislative and executive bodies should promote further research and create innovative regulations, such as “regulatory sandboxes”, permitting the deployment of automated minibuses (and AVs in general) without a safety operator, the safety operator being replaced by a (human) remote supervision that monitors several automated minibuses in parallel and intervenes if/when necessary.

Enabling the remote supervision of automated minibuses and vehicles (e.g. without a safety operator located inside the vehicle) and creating a dedicated legal framework are crucial steps for PTOs to deploy automated minibuses commercially at large scale and to provide economically viable transportation solutions in the future.

Some European Member States such as France and Germany, as well as the European Implementing (European Commission, 2022) and Delegated Regulations (European Commission, 2022) issued in July and August 2022, as well as the interpretation report published in February 2024 (EU Regulation 2022/1426), are paving the way from a legal point of view in Europe, by permitting experimentations and/or by enacting legal frameworks allowing remote supervision at distance under an ordinary approval regime. In February 2023, the Law Commission of England has published its advice related to remote driving for drivers beyond line of sight (Lawcom, n.d.), which includes the question of remotely driving AVs.

2.3.2 Qualifications of the “Safety Operators”

In parallel and during a “transition phase”, the qualifications currently expected from safety operators located inside automated minibuses are often as high or even higher than for drivers of conventional public transport vehicles (Roedl, n.d.), generating important functioning costs. Therefore, the requirements for safety operators should be watered down and adapted to the reality of operating automated minibuses. After a first period of practice where the AVs’ operational design domain is tailored to the physical location of their deployment, and problematic “hotspots” get identified, AVs are able to circulate with operation time over 90% in automated mode, where the automated driving system is successfully performing the automated driving task without any intervention of the safety operator.

The situation in Europe is quite behind in comparison to other countries. In 2022, Cruise and Waymo received commercial licenses to deploy small fleets of robotaxis in San Francisco, operating only with remote supervision: robotaxis are cruising the streets without any human driver inside, besides the passengers (TheLastDriverLicenseHolder, n.d.). Beginning in March 2024, Waymo received an authorisation to expand its robotaxis beyond San Francisco and south of the city to the Bay Area, as well as in parts of Los Angeles (TheLastDriverLicenseHolder, 2024c).

Other important topics shall be covered, as mentioned within the AVENUE research, for instance, as regards the passive safety of automated minibuses. Their particular (and new) architecture and features require urgently the definition of passive safety requirements. Reflexions and studies are necessary in order to arbitrate on important trade-offs in this field.

2.4 Recommendations Related to the Regulation of AV Data

2.4.1 Competition

The European Commission has published various legal acts and drafts related to current data issues, such as the Data Act and the Data Governance Act (amongst others). These acts are a positive first step, but further specific acts dedicated to the automotive industry, and automated vehicles, should be enacted quickly. Governments and lawmakers should provide their best efforts in creating legislations that anticipate and guide the technological developments. It should be avoided that, as seen until recently in the field of digitalisation, one or a few players get into a position where they can distort competition, due to their actual power on such market, generated by creating “winner takes it all” and “lock-in” effects within a nearly closed ecosystem (see details in chapter below about the vision of AVENUE). Besides open data (see below), open API and even open map are key issues to keep competition but also to better satisfy the citizen needs and to enable wealth (see vision).

The AVENUE project has demonstrated that vehicle manufacturers and/or tech companies developing automated vehicles and minibuses benefit of a “de facto” control over vehicle data. This means that they have control and can monitor key steps related to the collection of car data. They may, for instance, decide on the collection (for instance, sensors such as video cameras recording the AVs’ surroundings), the processing (storage of the data collected on temporary or permanent data storage devices), the transmission of data to third parties, the encryption for data etc.

Thus, vehicle manufacturers can decide, through agreements with other parties, to whom and to which type of data they provide access to. Such a “de facto” position places them in an undesirable gatekeeper position that should be mitigated (Ernst & Young, 2021; Andrasko et al., 2021) by creating a legal framework regulating data-related issues, as contractual arrangements between operators (in a form of self-governance) will not be sufficient.

One of the suggestions of the AVENUE project is to define through dedicated regulations open datasets according to the rules of the Data Governance Act that should be freely shared between the stakeholders, as well as open application programming interfaces to enable the transmission of data and leverage it for the greater public interest (Roedl, n.d.). The trustworthy mobility data-sharing system should accordingly create wealth for the society and provide control to the citizen and trust to the companies. The availability of mobility data also raises the question of privacy (see below) but also of competition as foreign competitor could use AI technology and less trustworthy data to take competitive advantage on European OEM (Original Equipment Manufacturer) competitors. A symmetry in competition should accordingly be guaranteed for European OEMs.

At international level, mutual assistance and questions related to data sovereignty should also be discussed, amongst others, as the data flow might imply that data collected from an automated vehicle located in country A is transmitted, processed and stored in another country B, sometimes outside of the European Union. This will cause problems when personal data gets transmitted to jurisdictions that do not have an adequate level of personal data protection.

2.4.2 Data Privacy

The deployment of AVs will raise many privacy concerns, as they are able to collect, store and transmit large amounts of personal data through their various sensors and communication channels.

Here are some of the key privacy issues posed by automated vehicles:

  • Definition of personal data: In the context of automated minibuses and mobility-related applications, what data is to be considered personal and non-personal (technical data) is not straightforward in all cases. Further guidance in this area would be necessary to avoid the proliferation of different interpretations of this concept, which has an important impact in practice, for example, on the anonymization of personal data (Campmas et al., 2021; Benyahya et al., 2022).

  • Processing of personal data: AVs can collect and store a large amount of personal data that can be used for various purposes, such as research and quality improvement, but also for marketing and establishing data profiles for commercial purposes. This may conflict with important data protection principles such as the data minimisation and the purpose limitation principles (Balboni et al., 2020; Finck & Biega, 2021).

  • Transparency: AVs should be transparent about the data they collect, store and share, so that the passengers are aware of the data that is being collected and how it is being used (Horizon 2020 Commission Expert Group, 2020). Achieving transparency towards third parties located outside AVs such as other drivers, cyclists and pedestrians will be harder to achieve, but solutions might be achieved through further research and discussion.

2.4.3 Cybersecurity

The storage and transmission of both personal and vehicle data must be protected to prevent unauthorised access and use of the data. This requires strong data security measures to ensure the protection of the data but also to prevent third parties to take control of AVs (see chapters above).

It is important for governments and manufacturers of automated vehicles to address these privacy issues (Vellinga, 2022), to build public trust in the technology and ensure that the personal data of the vehicle’s occupants is protected. Adequate privacy protections will help promote the widespread adoption of automated vehicles and ensure that they are used in a responsible and ethical manner.

2.4.4 Liability

The general opinion amongst legal scholars (Lohmann & Müller-Chen, 2017) today considers that automated vehicles in their current stage of deployment will not create major gaps in the applicable European liability framework, amongst others, due to the Motor Vehicle Insurance Directive, which requires that all vehicles registered in the EU hold mandatory third-party liability insurance to cover civil liability for the use of motor vehicles. However, various potential issues have been identified.

It should be avoided to place the burden of proof with the victim of an incident (as it might be the case in product liability cases, for instance (Piantoni et al., 2021; Buiten et al., 2021)). Due to the complexity of AI products and their algorithms, liability claims would become difficult or overly costly to prove for victims, and consequently they may not be adequately compensated. The same applies to (complex) decision-making processes, where the decision-making chain of AI products cannot be explained afterwards and how the (problematic) output resulted (Horizon 2020 Commission Expert Group, 2020).

Information asymmetry is another issue. As mentioned in the chapter dedicated to data, vehicle manufacturers currently benefit of a “de facto” control over the data collected and processed by the sensors of automated vehicles. To understand the liability chain (and causality between a defect and a damage), accessing vehicle data will be of paramount importance. Access to this data, and understanding the data, will be challenging for victims and their representatives.

One solution would lie in the reversal of the burden of proof, or in lowering the burden of proof, as suggested in the current revision of the Product Liability Directive. The creation of a trusted third party like it is existing in the automotive industry could be a solution (see vision chapter below).

In addition, the deployment of automated vehicles and minibuses implies a complex ecosystem with a plurality of actors involved in the services. This could result in increasing difficulties to assess where a potential damage originated and to assign liability (European Commission, 2020b). Forms of shared responsibility, as well as a shift towards liability of vehicle manufacturers and tech companies etc., should be studied, discussed and envisaged.

3 Accessibility

In the field of automated minibuses, the applicable accessibility requirements could still be improved by vehicle manufacturers, as demonstrated by the existing vehicle constructions, amongst others, in the AVENUE project. Moreover, the mobile applications related to the services provided by automated vehicles and minibuses also lack accessibility.

On the one side, deployers of automated minibuses, public transport operators and public transport authorities should include accessibility as an important criterion in the tender process. In many countries, accessibility is a prerequisite for the funding of vehicles by public authorities.

On the other side, vehicle manufacturers shall apply accessibility requirements defined by European standards in the earliest construction and design phase of automated vehicles (European Commission, 2018) (“accessibility by design” in a similar way than “privacy by design and by default” for data privacy). As a general principle, discriminatory design and service provision should be prevented (Horizon 2020 Commission Expert Group, 2020).

Accessibility of future automated public transport of all types needs to be thoroughly researched, and findings need to be designed into guidelines issued on public transport accessibility levels. Persons with reduced mobility (PRMs) need to become Cooperative Intelligent Transport Systems (C-ITS) nodes for emerging automated vehicles to recognise and protect them, given their potentially lower perceptibility and higher vulnerability. The impact of each new technology on PRM accessibility must be assessed during its design phase, and appropriate adaptability/personalisation means and strategies should be supported.

There is also a need for standardised and enforceable European legislation on accessibility across the mobility chain, as no such dedicated legislation exists for multimodal transport, be it for long-distance or local transport. The European sectorial legislation on air, maritime, railway and road transport does not fulfil this goal, as it (1) allows for exceptions that were largely used by Member States, (2) mainly focuses on long-distance transport and (3) does not support a wide range of the actual requirements of PRMs. The current legislation does not cover the possibilities of spontaneous travel (European Commission, 2018), i.e. there is a duty of prior notification in case of assistance requirements (Regulation on Rail Passengers Rights, n.d.), and provides inadequate support during travel and compensations for the loss or damage to accessibility equipment (for instance, in the cases of denial of travel and limited lost property compensations).

4 Conclusion

AVs might provide disruptive changes to the mobility of the twenty-first century. The regulatory environment in which they are tested and deployed can be a major incentive (International Association of Public Transport, 2021) or an obstacle to their development. The governance decides also about the citizen centricity, the creation of wealth but also the business viability in particular the creation or not of closed ecosystems. European executive and legislative bodies are now beginning to understand the need for AV regulations and the role they can play. The integration of AV in MaaS/ITS is a new challenge as well (see vision below). Encouraging signals have been emitted recently with the adoption of innovative regulations (Roedl & Partner, 2022). However, much legal work remains ahead, in order to avoid having regulations that are too far behind the technological developments, opening a regulatory “no man’s land” that would be beneficial for a few private stakeholders only, and not to society as a whole.