Autonomous vehicles under pressure: dependence on infrastructure and the challenges of regulation and civil liability
Intelligent transport systems and connected and automated mobility are now a reality. Autonomous vehicles [1] are an inescapable part of this new state of affairs.
Autonomous vehicles at almost every level promise profound changes in mobility and significant gains in safety and efficiency.
There is, however, a web of technological and infrastructural dependencies and legal challenges that need to be overcome.
Among others, the safe integration of this technology requires collaboration between reliable AI systems, extensive testing, robust digital networks, smart sensors, smart roads and funding.
In addition, there is still a need to create clear rules on liability as the model shifts focus from the human driver to the ecosystem formed by manufacturers, operators, owners, insurers and infrastructure authorities.
Cooperation between technology, law and urban planning is essential to alleviate this pressure on autonomous vehicles.
Technological and infrastructural dependence
The safety of autonomous vehicles was once believed to be ensured exclusively by V2V technology (i.e. by sensors and vehicle-to-vehicle communication systems).
This, however, is not the case. This safety is inseparable from V2X (Vehicle-to-Everything) connectivity, which necessitates the existence of excellent communication support between vehicles and, above all, connection between the vehicle and the road infrastructure itself, through mobile networks and data platforms. All of this must exist in real time. A delay in communication can compromise safety as the reaction required to avoid collisions requires an automatic response in milliseconds.
Alongside connectivity, the success of these projects depends on smart infrastructure. Smart roads include road sensors, cameras, and other monitoring systems to create a "digital twin" of both the road and the road environment, providing real-time data on traffic and obstacles. The effectiveness of these systems depends on the quality and maintenance of the road. It also depends on the sensors and the accuracy, timeliness and quality of the data collected.
Implementing such complex urban environments will come at extremely high costs. This type of road requires strategic urban planning, and the current landscape and design of roads will require significant physical and digital adaptations.
Although 5G and cooperative systems promise to accelerate the process, there is still a long way to go in terms of technological and legislative maturity.
Projects such as C-ITS [1] , 5G corridors [2] and public-private cooperation platforms seek to create the telecommunications and cooperation conditions necessary for connected and automated mobility. But we are still a long way off.
Legal changes
Legal challenges will also need to be addressed and it is essential to define who should be liable for accidents involving autonomous vehicles.
The dissociation between control (algorithmic) and responsibility (traditionally human) is therefore the core dilemma.
The trend is to transfer responsibility from the driver to manufacturers [3] , software developers [4] and the system [5] without holding operators completely non-liable since they must always maintain certain levels of responsibility [6] .
Responsibility should vary or be shared depending on the level and use of automation. In level-0-to-2 vehicles, the human is solely responsible for driving. In level-3 vehicles, the system is already driving, although it can ask the human to take control at any time. In this case, responsibility is shared and depends on who is in control at the time of the accident. If the human does not take control when asked, the human will be responsible.
In level-4 and 5 vehicles, the system is already driving. In this case, responsibility tends to fall on the manufacturer, software developers or a fleet operator.
Without prejudice, users and owners must remain responsible for breaching their duty to use the system properly in accordance with instructions and maintenance requirements.
Responsibility may also fall on infrastructure, whether road authorities or concessionaires, when failures in digital or physical infrastructure cause accidents (e.g. faulty sensors or poorly maintained digital networks).
Special attention is also required in terms of insurance. In the EU, motor insurance for third-party damage is compulsory, but automation increases the need for additional coverage by manufacturers and technology operators, given the growing technical nature of risks and potential infrastructure failures.
Evidence and investigation problems
Determining the causes of accidents is complex. Among other things, it involves capturing and analysing evidence relating to algorithms, sensors, communication data, interactions between systems, etc. To this end, data recording systems, similar to aircraft black boxes, may need to be installed.
Rules need to be created to mitigate the technological asymmetry between manufacturers and victims in the access to technical data. Since manufacturers control this data, victims could have difficulty proving their cases. Reversing the burden of proof could make this easier.
Finally, there are social and regulatory pressures. Every accident involving autonomous vehicles appearing in the media increases scrutiny. Autonomous vehicles are expected to be safer than humans, not simply “as good as” human drivers. Even rare failures have a major impact on public perception and reputational risk is high.
Conclusion
The pressure on autonomous vehicles is not only technological. Pressure stems from the interdependence between reliable connectivity, smart roads, clear liability rules and social acceptance. Without collaboration integrating these factors, the promise of safety and efficiency risks falling short of expectations.
