AI is revolutionizing transportation by making it more sustainable. This application in autonomous vehicles has its own set of complexities concerning liability in case of infractions. The methodology employed in this study involves a comparative legal analysis approach. This includes a comprehensive analysis of primary legal documents to understand the current legal landscape in the selected jurisdictions. Additionally, the study draws on a real-world comparative analysis and examines liability claims to gain practical insights into the legal complexities. Secondary sources include academic literature, industry reports and news articles. This paper examines various aspects of criminal responsibility of AIbased AVs, drawing comparisons among US, Germany, UK, China and India. The rationale for comparing these countries lies in their diverse legal frameworks. These countries were chosen for their technological advancements and contrasting regulatory approaches to liability in AI-enabled autonomous vehicles.
The goal is to compare how different countries have approached this problem by analyzing their legal frameworks and responses to it. It explores various approaches for ascertaining human errors that result in crime, such as intervention or moral agency on the part of AI, and identification of the primary offenders in incidents involving AVs. However, it shows that every country has its own unique way within its respective jurisdiction. For instance, India and USA have a loose interweaving network of state laws, while UK made a pioneering piece of legislation in 2018 called the Automated and Electric Vehicles Act, 2018. Germany applies strict safety standards and distinguishes liability based on the operating mode of the vehicle. Contrarily, China also aims to establish a very strict liability regime for AVs. Lastly, as an outcome of this study, it was found that there is a pressing need for globally agreed upon legal standards to encourage technological advancements, ensuring there is innovation invoking minimum risk.
Self driving cars are the kind of vehicles that employ artificial intelligence as well as other sensors to operate autonomously. This is done without human intervention or supervision.
These vehicles can make transportation safer and more efficient by helping people get where they are going easier, in an environmentally sustainable manner. 1 AVs 2 also have significant legal and ethical connotations, especially criminal liability, since they pose questions about who is at fault for their actions and outcomes. However, the shift from conventional vehicles to AVs is neither consistent nor continues in a straight line; rather, it progresses in degrees with different levels of automation, capabilities and applications. The SAE 3 has come up with six levels of automation for on-road vehicles. 4 They are as follows:
Level 0: No Automation-The human driver performs all driving tasks.
Level 1: Driver Assistance-It is specified that the vehicle system shall be able to provide assistance to the human operator with either steering or acceleration/deceleration. Nonetheless, it is clearly stated that these two cannot be offered simultaneously.
Level 2: Partial Automation-The automotive system assists the human driver in steering and acceleration/deceleration. However, the driver must always be ready to take control of the vehicle.
Level 3: Conditional Automation-The vehicle can drive by itself under some conditions. However, a human being must be ready to take over control at any moment, should there be a need for assistance from him/her or a fault happens in the system.
Level 4: Elevated Degree of Automation-These are systems that perform all driving tasks under specific conditions without requiring participation from humans or supervision by drivers, like in a traditional car.
Level 5: Complete Automation-The system performs all driving tasks under all circumstances. In no way does it involve a person who is expected to participate in whatever form.
Indeed, the issue of criminal liability is a significant and contentious aspect within the legal framework for AVs or their AI 5 systems. Criminal liability pertains to the legal obligation incurred for the commission of a crime, which could lead to sanctions. It includes fine, imprisonment or other penalties. 6 The primary elements consist of the two: i.
Actus reus (the wrongful act) and ii.
Mens rea (the guilty mind). 7 However, the process of integrating these principles into the operations is neither direct nor unequivocal. The inherent complexities of these advanced technologies pose significant challenges to the unambiguous incorporation of existing legal elements. According to some scholars, there are three basic models to cope with this phenomenon within the current definitions of criminal law. 8 These models are:
a. The Perpetration-by-Another Liability Model: In this model, a human being who uses, controls or directs the AI system that commits a crime will be held liable as the perpetrator, whereas the AI system is seen as an instrumental tool.
b. The Natural Probable Consequence Liability Model: This model holds that if someone creates, programs or allows an autonomous AI system to commit a crime, then they are liable as an accessory since it was foreseeable that such criminal conduct would naturally flow from their acts or omissions.
c. The Direct Liability Model: This model holds that where an AI system can be said to have had enough autonomy, intelligence and moral agency, it may also be considered criminally responsible as a separate legal person.
These models have different benefits and drawbacks and thus, they may be used differently depending on the degree of automation, capacity as well as purpose for which the AI system is put in place. Some legal systems could blend various models or even create others to respond specifically to challenges posed by AI systems.
A primary obstacle in the regulation of these vehicles is the identification of the party to be held legally accountable upon any violation. 9 The question of who bears legal responsibility for acts done by AV remains unanswered. Liability could be attributed to either humans operating them, AI system running them, vehicle manufacturers, software developers making programs that make these cars function, service provider or a combination of these parties. This issue can be approached differently in different jurisdictions depending on their underlying principles, values and objectives. Various legal viewpoints on the criminal liability of autonomous vehicles within the jurisdiction of India and other nations are discussed hereunder. 10
India is a common law jurisdiction following the principles of tort law. It includes the concept of negligence, strict liability and product liability as well. India currently lacks any specific legislation addressing autonomous vehicles. Motor Vehicles Act, 1988 is the main ruling that governs road transport & traffic in India. The MVA 11 defines a
This content is AI-processed based on open access ArXiv data.
