Document Type : Original Article
Author
Assistant Professor, Faculty of Humanities, Islamic Azad University: Qazvin Branch, Qazvin, Iran. keshavarzsmaeil@gmail.com
Abstract
Introduction
The development of computer science since the mid-twentieth century has led to the formation of the phenomenon of artificial intelligence. Today, artificial intelligence, by using precise calculations, access to very large volumes of information, and a lower error rate, has demonstrated its superiority over human performance in thinking and decision-making in many fields. However, this development has also brought risks with it. One of these risks is the issue of compensation for damages resulting from the behavior of artificial intelligence systems, as well as the basis for civil liability resulting from their harmful actions. The emergence of damages caused by artificial intelligence systems has led to the tendency of lawyers and legislators in various countries to establish special regulations regarding human factors related to artificial intelligence activities, from the production and manufacturing process to its use in relationships between humans. The European Union is one of the leading regions in regulating civil liability arising from the operation of AI. This article seeks to comprehensively examine various EU documents separately in the field of civil liability, factors related to the performance of artificial intelligence systems, and compare the current Iranian legal system with EU documents. In the final step, a proposed solution to address the shortcomings and criticisms of EU documents as well as the Iranian legal system will be examined.
Methods
The main research method in this study is a descriptive-analytical method. The sources required in this study have been collected through library methods and by referring to reliable databases and websites. The main goal of this study is to determine the basis for civil liability of various factors related to the performance of artificial intelligence systems. In this context, two legal systems of the European Union and Iran were selected. Regarding the European Union, various documents were studied, such as the 2020 European Parliament Resolution, the 2022 PLD Proposed Directive on Defective Products, as well as the 2022 AILD Proposed Directive. Also, despite the lack of a specific law in Iranian law regarding civil liability arising from the performance of artificial intelligence, the capacities of related laws such as the Consumer Protection Law and the E-Commerce Law to respond to the harmful performance of artificial intelligence were examined.
Results and Discussions
European Union documents are one of the leading legal documents in the field of regulation regarding the obligations, requirements and responsibilities of various factors related to the performance of artificial intelligence systems. In short, Resolution 2020, PLD's proposal and instructions for artificial intelligence producers, contains a no-fault liability system. But regarding providers and users of artificial intelligence AILD Directive And the artificial intelligence Act 2024 is proportional to the degree of risk of artificial intelligence systems, including liability based on fault (presumed fault for high-risk systems). However, due to the differences between documents regarding the basis of civil liability of various factors, as well as the interference and mixing in the definition of some titles such as producer, operator and provider, the preparation of a single recipe for the precise definition of each of the titles and determining the basis of civil liability of each one of factors are suggested separately to resolve the ambiguity and complexity of these documents. Regarding the laws of Iran, the current laws do not have the ability to respond to the challenges and harms of the emerging phenomenon of artificial intelligence and the progress of this technology and the degrees of danger and harm they cause, and the need to enact a comprehensive law to determine the legal framework and civil responsibility of each of the factors related to the performance of artificial intelligence appears to be an urgent matter. Until then, the provisions of the Consumer Protection Act And the Law of Electronic Commerce and Note 2 of Article 526 of the Islamic Punishment Act of Iran will rule regarding the harmful behavior of artificial intelligence systems, that due to some conflicts and deficiencies, the amendment of the conflicting articles and the addition of amendments to the mentioned laws in order to properly cover the types and Various degrees of high-risk and low-risk artificial intelligence systems will be essential. In summary, in the field of Iranian law, the author suggests two suggestions: 1- The most appropriate solution is to regulate the rules governing the civil liability of artificial intelligence systems in a specific manner. The 2024 Artificial Intelligence Law and its instructions can be a useful and appropriate guide for establishing regulations in this field. 2- The second suggestion (in case the first proposal cannot be implemented), regarding the civil liability of manufacturers (including suppliers) of artificial intelligence systems, is to eliminate the conflict between the note to Article 2 and Articles 18 and 19 of the Consumer Protection Law by amending the note to Article 2 and removing the condition that the supplier be aware of the defect in production, in order to establish a system of liability without fault for manufacturers and suppliers of artificial intelligence systems.
Conclusion
The 2020 Resolution, the PLD Proposal 2022 for AI Producers contain a system of no-fault (strict) liability. However, for AI Providers and Deployers, the AILD Directive and the AI Act 2024 contain, in proportion to the degree of risk of AI systems, presumed fault-based liability for high-risk systems and fault-based liability for low-risk systems. In the current Iranian legal system, the Consumer Protection Law can be considered to include no-fault liability of the manufacturer (including the supplier) of defective artificial intelligence systems by removing the note in Article 2 of this law. With respect to users of artificial intelligence systems, Article 78 of the Electronic Commerce Law and Note 2 of Article 526 of the Islamic Penal Code cannot, in the current situation, respond to the risk levels of artificial intelligence systems
Highlights
- The liability of the AI producer in the PLD is no-fault liability.
- According to Article 3 and 4 of proposal AILD 28.09.2022, the liability of the AI provider and deployer is presumed fault-based liability (Rebuttable Presumption of causation) for high-risk systems and fault-based liability for low-risk systems.
- Iranian law requires the establishment of a specific law for the harmful operation of AI.
Keywords
Main Subjects
)