The responsibility assessment in the scenario of driving AI-based vehicles

To evaluate behaviours involving autonomous artificial intelligence, it is necessary to consider them legitimate behaviours. The most effective way to prevent illegal activities is through legal acts, as the judiciary is largely unfamiliar with AI and the research of such systems requires a high level of specialised knowledge. Compliance with laws and regulations should serve as the criterion for judging the legitimacy of such behaviour. Therefore, legislators and competent authorities must complete relevant laws on AI and establish its development as a priority in national policies. The characterisation of legal norms, particularly the legal obligations of developers, is a regulatory gap that urgently needs to be addressed. Until laws are in place, the application of AI is challenging. Nevertheless, once relevant laws are established, the autonomous systems, such as self-driving cars or security machines, will be able to operate legally and without the risk of committing a crime which shall lead to the development of administrative procedures in case of incidents.

Conclusions

The development and use of autonomous artificial intelligence have brought about significant changes in various industries, particularly in the automotive industry with the introduction of selfdriving cars. While the benefits of AI are undeniable, there are also potential risks and uncertainties associated with its operation. Developers and users must comply with laws and regulations that govern the development and use of AI to ensure that their actions are legal and consistent with societal norms. Furthermore, it is essential to establish a legal framework that promotes responsible and ethical development and use of AI rather than relying on economic or experiential measures of benefit. In the event of accidents involving autonomous AI, it is important to distinguish whether the system has autonomy or not and to what extent to determine who should be held responsible. Behaviour that violates the law is considered illegal and is against the legal order of the state. However, legitimacy is not just the absence of illegality. Whether a behaviour is illegal, we need to evaluate it against the norms of the entire legal system, not just, for instance, administrative and criminal laws, and to observe whether a behaviour in the form of an action violates the legal order as a whole. In a society that values the rule of law, in principle, behaviour compliant with the law shall not be punished. When a behaviour is based on the law, it conforms to legal norms, does not disrupt the legal order, and cannot be deemed illegal.

The behaviour according to the law is not unlimited. Once the behaviour exceeds the legal norms with a reasonable expectation, it becomes illegal. It is noteworthy that in criminal law, there are regulations for reducing and exempting penalties for legitimate defence. In addition, in individual cases, acts following the law should be judged based on whether their substance is justified, even if they have legal forms. Although administrative rules do not have a direct external effect as legal norms, they have an indirect and de facto external effect.

The responsibility for accidents caused by autonomous AI systems will depend on the circumstances (in August 2017 the ethics Commissioner of the German Federal Ministry of Transport and Digital Infrastructure Council released a set of 20 ethical guidelines for driverless vehicles where paras 6 and 9 states that in cases of unavoidable collisions, human life must take priority over animals and property.) and the behaviour of the user and developers involved. Responsibility for the behaviour of users and developers is critical for ensuring the safe operation. When a selfdriving car is in use, the behaviour of the user is critical in ensuring the safety of the vehicle. If the autonomous system causes an accident due to the user's negligence, the user will be held responsible. However, if the autonomous system causes an accident due to a defect in the system, the developer will be held responsible. It is still a relatively new concept to hold autonomous systems accountable for illegal actions. While fully autonomous systems may eventually emerge, current technology and law theory still place the object of legal evaluation on human behaviour with the essence of wrongfulness. It is vital to strike a balance between the advantages and disadvantages of self-driving cars. However, if this is not possible, it is better to forecast and prevent the negative consequences first and reap the benefits later. The concept of permissible risk, which balances the benefits and disadvantages of risky behaviour, may not be easily applicable to the development of offensive autonomous systems. The lack of established standards and habits for measuring a ratio makes it difficult to judge due to the tolerance for risks.

The behaviour of autonomous AI systems is based on algorithms designed by developers, which means that risk control measures must be put in place during the development stage. Additionally, the legal issues of autonomous AI systems, especially in terms of the behaviour of the developer or restrained, may not be fully resolved by the elements of appropriateness. Therefore, it is prudent to exercise caution and avoid any behaviour that may be socially or legally harmful. While statutory regulations are extraordinarily legitimate acts that can serve as grounds to prevent violations, it is essential to evaluate behaviour based on whether its substance is justified, even if it has legal forms.

Real-world testing is necessary to fully realize the potential of unmanned vehicles, and careful planning and management are crucial. Self-driving cars hold the most promise, but more work is needed to make them a viable option for daily transportation. It is essential to understand the different types of unmanned vehicles and their implications for accidents and liability. Developers must comply with relevant laws and regulations and bear legal responsibility if they fail to do so. Legislative measures should distinguish between unmanned vehicle experiments with remote control and automatic operation, and developers who comply with the law should not be held accountable for accidents or injuries that occur during experiments involving autonomous intelligence unless there is a violation of the approved experimental plan. Prosperity in the development of selfdriving cars depends on the encouragement of talent and funding through relevant laws and regulations, leading to the integration of AI into daily life.

Moral responsibility can exist without legal responsibility, and legal responsibility can reinforce our moral judgments. However, legal responsibility does not necessarily depend on a moral one, and for an action to be legally culpable, it must also be morally reprehensible. In the case of self-driving cars, factors such as morality and collision ethics play a significant role in decisionmaking. As the sensors become more sensitive, the system's designers must consider that the factors used to prioritize human life are morally justifiable. The legal punishment for self-driving cars must entertain the ethical contemplations that went into their design, not just the fact that an accident occurred. Therefore, it is crucial to think carefully about the implications of collision ethics and the impact of different characteristics on decision-making in self-driving cars. Finally, the supreme goal of designers should be to create machines that can assist in decision-making while adhering to ethical and moral principles.

References

1. Bohm, Felicia, and Klara Hager. (2015). Introduction of Autonomous Vehicles in the Swedish Traffic System: Effects and Changes Due to the New Self-Driving Car Technology. Uppsala universitet, Fasta tillstandets fysik, pp. 17-18.

2. Burri, Susanne. (2022). Conceptualising Morally Permissible Risk Imposition Without Quantified Individual Risks. Synthese (Dordrecht), vol. 200, no. 5, pp. 415-415

3. Coeckelbergh, Mark. (2016). Responsibility and the Moral Phenomenology of Using Self-Driving Cars. Applied Artificial Intelligence, vol. 30, no. 8, pp. 748-57

4. Coglianese, Cary, and David Lehr. (2017). Regulating by Robot: Administrative Decision Making in the Machine-Learning Era. The Georgetown Law Journal, vol. 105, no. 5, p. 1182.

5. De Bruyne, Jan, and Jarich Werbrouck. (2018). Merging Self-Driving Cars with the Law. The Computer Law and Security Report, vol. 34, no. 5, p. 1153

6. Eggers, Felix, and Fabian Eggers. (2022). Drivers of Autonomous Vehicles-Analysing Consumer Preferences for SelfDriving Car Brand Extensions. Marketing Letters, vol. 33, no. 1, p. 103

7. Gill, Tripat. (2020). Blame It on the Self-Driving Car: How Autonomous Vehicles Can Alter Consumer Morality. The Journal of Consumer Research, vol. 47, no. 2, p. 280

8. Konig, M., and L. Neumayr. (2017). Users' Resistance Towards Radical Innovations: The Case of the Self-Driving Car. Transportation Research. Part F, Traffic Psychology and Behaviour, vol. 44, p. 46

9. Nidhi Kalra, James M. Anderson & M. Wachs. (2009). Liability and Regulation of Autonomous Vehicle Technologies, RAND Corporation, pp. 6-7

10. Norm Violation as the Criterion for Objective Attribution. Should the Assumptions of Increased Risk Theory Be Rejected? Studia Prawno-Ekonomiczne, vol. 114, 2020

11. Pranam, Aswin. (2018). Product Management Essentials Tools and Techniques for Becoming an Effective Technical

12. Product Manager. 1st ed., Apress, p. 140.

13. Ross, Hans-Leo. (2021). Safety for Future Transport and Mobility. 1st ed., Springer

14. Salmanova, O.Yu, and A.T. Komziuk. (2021). Administrative Liability for Violating the Rules of Stopping, Parking, Vehicle Parking. Bulletin of Kharkiv National University of Internal Affairs, vol. 94, no. 3, pp. 150-61

15. Salvendy, Gavriel, and June Wei. (2020). Human- Computer Driving Collaborative Control System for Curve Driving. In Design, Operation, and Evaluation of Mobile Communications, vol. 12216, Springer International Publishing AG, pp. 1-15

16. Tao, Chuanqi, Gao, Jerry & Wang, Tiexin. (2019). Testing and Quality Validation for AI Software-Perspectives, Issues, and Practices, IEEE access vol. 7, pp. 120164-120175

17. Watzenig, Daniel., and Martin, Horn. (2017). Automated Driving: Safer and More Efficient Driving in the Future. Edited by Daniel. Watzenig and Martin. Horn, 1st ed., Springer International Publishing

Размещено на Allbest.ru