Text | Brother Octopus
Editor|Beachburg
I. Typical cases and legal norms of autonomous vehicle traffic accidents
(1) Typical cases
Typical case 1: Tesla car accident
In 2019, there was a Tesla sedan accident, and the driver of the Tesla sedan failed to recognize the traffic light at the intersection and hit another car and killed two people because he used Tesla's Autopilot system.
In the case, the driver was charged with manslaughter.
Typical case 2: Uber self-driving car accident
In March 2018, Uber's self-driving car hit a pedestrian while driving, resulting in their death. Prosecutors determined that the accident was caused by the driver's negligence and Uber was not criminally liable.
The findings said that human error caused the crash, but suggested that Uber and the driver share responsibility for the aftermath of the accident.
Through these two typical cases, it can be seen that traffic accidents caused by self-driving cars cannot be underestimated and must be treated with caution. The controversy reflected behind it is also self-evident, solving the current development dilemma.
In the face of how to resolve the problem of liability determination caused by autonomous vehicles, we must give a feasible solution path from a legal perspective to promote the healthy development of autonomous vehicles.
(2) Relevant legal norms
Self-driving cars are the product of a new round of scientific and technological revolution, and it is necessary to understand the relevant regulations for self-driving cars from the normative level.
Through review and collection, the main foreign normative requirements for autonomous vehicles are the United States, the United Kingdom, Germany, Japan, Austria, and Singapore.
1. Foreign legal norms
In 2016, the United States issued the Federal Autonomous Vehicle Policy Guidelines, which regulate how autonomous driving systems detect obstacles and other aspects.
In the United Kingdom, the Autonomous and Electric Vehicles Act was published in 2018, and the driver will not be held accountable for traffic accidents that occur while the driver is in autonomous driving mode, and the liability will be borne by the autonomous driving entity.
In the UK, the responsibility holder after the autonomous driving mode is turned on has been clarified.
Through the combing of the legal norms of the United States and the United Kingdom, it is not difficult to find that the two countries provide a basis for the development of autonomous vehicles through institutional design, which involves provisions on automatic driving systems and responsibility.
This also reflects the difficulties in determining criminal liability caused by safety accidents caused by autonomous vehicles in practice.
Germany amended the Road Traffic Safety Law to clarify the applicable product liability for autonomous driving, and the liability for accidents caused by it was accordingly transformed into product liability.
In addition, subsequent bills also clarify the applicable product liability for autonomous driving, which provides guidance for exploring liability issues arising from autonomous vehicle traffic accidents.
At the same time, German law also stipulates that if the accident occurs during the autonomous driving phase, the car manufacturer is liable.
By specifying the division of responsibilities in detail, it is clear that the responsibility is borne by the driver and the automobile manufacturer for different situations, and provides an institutional basis for exploring the issue of determining criminal liability.
In addition, it stipulates that the manufacturer shall be liable in the event of an accident caused by an obvious failure such as incorrect operation of the system.
By combing through the legal provisions of Germany and Japan on self-driving cars, it is not difficult to find that both countries have stipulated the division of responsibility for traffic accidents, making the subject of responsibility for traffic accidents caused by self-driving cars clearer.
Moreover, further comparative analysis, there are some differences in the division of responsibilities between the two countries.
Singapore also introduced regulations for autonomous vehicles on January 31, 2019, and its "Technical Specification 68" stipulates that autonomous vehicle developers and operators should operate a quality management system in accordance with best practice standards.
and an independent evaluation, the specification also makes it mandatory for autonomous vehicle developers to provide sufficient evidence that autonomous vehicles are safe and fit for their intended purpose.
It is not difficult to find out through the specification that Singapore refines the provisions that the evidence used to prove that self-driving cars are safe and fit for their intended purpose is provided by the developers of autonomous vehicles.
This provision provides a rule for dealing with difficult issues involving the determination of developer responsibility.
Austria's Vienna Convention on Road Traffic Safety regulates the situation in which autonomous vehicles are on the road, where the driver must supervise the car at all times and, in exceptional cases, the driver will take over the car.
Comparing the Singapore and Austrian regulations, we can see that both countries have detailed rules for the normal operation of autonomous vehicles.
2. Relevant regulations of the mainland
Regarding who should bear the responsibility for traffic accidents caused by self-driving cars, the Road Traffic Safety Law of the People's Republic of China on the mainland does not stipulate.
However, through the Shenzhen Special Economic Zone issued the "Shenzhen Special Economic Zone Intelligent Networked Vehicle Management Regulations", it is not difficult to find that the determination of responsibility for autonomous vehicles has been on the legal agenda.
The regulations issued by Shenzhen City stipulate three types. The regulation is mainly aimed at L3 autonomous vehicles.
The Regulations stipulate that if a traffic accident is caused by a defect in the quality of the car, the driver may recover compensation from the manufacturer or seller of the automobile after bearing the liability for damages according to law.
If an accident is caused by a quality defect in the automobile product, the driver can compensate the manufacturer and seller of the automobile for responsibility.
Through the combing of current laws and regulations, whether foreign or domestic, it involves self-driving cars, which is based on the development needs of the times, and it is also an institutional response to the current problems of self-driving cars.
The regulations involved all explain in detail the division of liability for traffic accidents caused by autonomous vehicles, providing institutional guidelines for studying this issue.
Through the further combing of the provisions, it can also be analyzed what difficulties exist in the determination of criminal liability of current autonomous vehicles, and also lay the foundation for analyzing their difficulties and further exploring a path for development.
Second, the dilemma of determining criminal liability for the crime of traffic accidents caused by L3 autonomous vehicles
The general theory of criminal law holds that what constitutes a crime is constituted by the system of criminal theory.
The actor, the subjective aspects of the actor and the causal relationship between the act and the damage are the three most basic elements of establishment.
This study analyzes the obstacles in the determination of criminal liability, subjective aspects and causality in the event of traffic accidents in L3 autonomous vehicles based on the characteristics of L3 autonomous vehicles, which lays the foundation for the theoretical analysis of the fourth part.
(1) The impact of autonomous driving on the main theory of criminal law
Legal liability is realized through legal sanctions, and the responsible subject is the prerequisite for assuming legal responsibility. If the subject is uncertain or doubtful, then the legal liability will be false.
The determination of the actor is crucial to the determination of liability in the event of a traffic accident caused by an L3 autonomous vehicle. But in Level 3 autonomous vehicles, the driver's position is reset compared to a conventional car.
When the criminal risk caused by L3 level autonomous vehicles, there is a danger that the subject of criminal liability will be blank.
1. Reset of the driver
According to the provisions of the current criminal law on the mainland, the scope of criminal liability is limited to natural persons and units. Specific to the crime of causing a traffic accident, according to Article 133 of the Criminal Law, the subject of responsibility can only be a natural person.
However, in Level 3 autonomous driving, it is difficult to identify the driver due to the intervention of the autonomous driving system.
The use of self-driving technology has to some extent turned human drivers into passengers. In the normal autonomous driving process, it is the autonomous driving system that really plays the role of "driver", not the human driver.
However, humans have not completely got rid of the "driver" role, and only in special circumstances do human drivers need to return to the "driver" role.
Humans and autonomous driving systems who are the drivers affect the responsibility for traffic accidents.
In particular, when the autopilot system does not give a takeover prompt or a human does not take over although a takeover prompt is issued, it is controversial who exactly is the driver.
The first view is that a human driver is a vehicle driver.
However, there is a view that this only applies to the determination of responsibility in the event of a traffic accident on public transportation after a human driver takes over a self-driving car.
At this stage, Continental's self-driving cars can only be tested on closed roads or venues, which do not belong to the field of public transportation, so they do not constitute the crime of traffic accidents.
The second view is that self-driving cars and human drivers together constitute two drivers.
The U.S. Department of Transportation document, Preparing for Future Transportation: Self-Driving Cars 3.0, expands the definition of driver and operator to make both the driver of the autonomous driving system and the person who turns on the autonomous driving mode the driver.
However, Continental's "Classification of Automobile Driving Automation" defines a human driver as a backup user, who can only become a driver in the traditional sense after taking over the car.
In short, L3 autonomous vehicles are not vehicle drivers when humans don't take over driving the car.
It can be seen that the reset of the driver's status has made it more theoretical obstacles to determining the crime of traffic accidents. The "human-machine escort" model of L3 autonomous vehicles also makes it more difficult to determine who is responsible for traffic accidents.
2. Blanking of the responsible entity
In the traditional driving mode, a subject with criminal responsibility can become the responsible subject of a traffic accident.
When self-driving cars with learning ability cause criminal risks, the current criminal law exposes its limitations in the determination of responsibility, and it is difficult to identify the subject of criminal responsibility.
The driving mode of self-driving cars has caused a huge impact on the traditional criminal law subject theory, and it is extremely difficult to determine the criminal subject.
In Level 3 autonomous vehicles, under normal circumstances, the real operator of the car is the automatic driving system, and the driver's status is largely blurred, so that the subject of criminal responsibility may "disappear".
First, when an L3 autonomous vehicle is involved in a traffic accident during autonomous driving, the driver cannot be the subject of responsibility for the traffic crime.
First, the driver's use of autonomous driving modes under the right conditions meets the characteristics and design requirements of autonomous vehicles, so it will be considered unreasonable for the driver to commit a traffic offence.
In the conventional driving mode, because the driver has the obligation to check the condition of the vehicle, failure to check the condition and driving behavior as a whole is evaluated as illegal driving.
However, in Level 3 autonomous driving, because the autonomous vehicle itself has autonomous detection and control functions, the driver's behavior of not checking the condition of the vehicle cannot create a danger alone, and it is difficult to determine that it constitutes a traffic accident.
Second, when Level 3 autonomous driving is autonomous, the driver essentially plays the role of using a human.
According to Article 7 of the "Connected Vehicle Management Specification", the driver's behavior does not meet the harmful behavior and criminal place requirements of the crime of traffic accident, and does not constitute the crime of traffic accident.
Some people even believe that the driver has neither control nor subjective fault, and if he bears criminal responsibility, it is contrary to the spirit and essence of the criminal law.
Taking a step back, even if the driver is considered to have committed a criminal act under criminal law, it is likely that the result will not be avoided due to the intervention of the autonomous driving system.
If a self-driving car suddenly changes lanes abnormally or suddenly fails to cause a traffic accident, it seems that the driver cannot be blamed for the crime of causing a traffic accident.
Finally, the responsibility of the driver for the crime of causing a traffic accident may violate the principle of proportionality of punishment. If only the driver is held criminally liable, it means that the user bears full duty of care at all times while the car is moving.
Second, autonomous driving systems also do not seem to be the subject of traffic crimes.
The existing criminal law and traffic legal systems are constructed with human drivers as the main body, and various liability regulations are also set up with human behavior as the center.
Specific to the crime of traffic accidents, the crime is set up in the traditional human driving model, and its premise is that human behavior is the center.
From this perspective, it is not possible to evaluate the conduct of L3 autonomous vehicles in traffic accidents within the existing criminal law framework. Moreover, the premise of this question is whether artificial intelligence legal subjects should be qualified.
Academic discussion on this issue is in full swing, but far from a consensus.
From this point of view, it is not difficult to find that the intervention of the automatic driving system has led to theoretical obstacles to attributing traffic accidents to the self-driving car itself or the driver, at least there is a need to re-clarify.