Next Article in Journal
Numerical Study of Topographic Effects on Wind-Driven Coastal Upwelling on the Scotian Shelf
Previous Article in Journal
The Symbiotic Relationship between the Antarctic Limpet, Nacella concinna, and Epibiont Coralline Algae
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Potential Liability Issues of AI-Based Embedded Software in Maritime Autonomous Surface Ships for Maritime Safety in the Korean Maritime Industry

College of Maritime Sciences, Korea Maritime & Ocean University, Busan 49112, Korea
*
Author to whom correspondence should be addressed.
J. Mar. Sci. Eng. 2022, 10(4), 498; https://doi.org/10.3390/jmse10040498
Submission received: 2 March 2022 / Revised: 11 March 2022 / Accepted: 14 March 2022 / Published: 3 April 2022
(This article belongs to the Section Ocean Engineering)

Abstract

:
Maritime Autonomous Surface Ships (MASS), an emerging area of digital advancement in shipping and shipbuilding industries, presents a different legal paradigm from that of existing ships. Existing maritime-related industries, including shipping, shipbuilding, and logistics, based on large hardware called ships, are rapidly changing into highly autonomous software-centered structures. This study is focused on analyzing the legal issues in preparation for MASS’ commercial operations in the future by applying comparative methods centered on the Republic of Korea and the United Kingdom. The study’s results contribute to the criteria for the manufacturing responsibility of autonomous ship-embedded software and a desirable legal policy improvement plan. Various legal issues and macro legal policy directions related to software product liability were identified. The study also presents concrete implementation strategies to achieve an ideological harmony and a balance between equitable damage relief and the advancement of related technologies to ensure maritime safety in the maritime industry. Based on the issues identified and their legal policy alternatives, it is hoped that the institutional ideal of product liability, which promotes technological advancement and protects consumer rights, is realized in the software domain as well.

1. Introduction

1.1. Background of the Study

For thousands of years, the fact that ships’ navigation is possible only when people are on board has remained unchanged. Currently, autonomous navigation with the Fourth Industrial Revolution technologies is gradually developing, and it is expected that fully unmanned ships will emerge in the near future [1]. The International Maritime Organization (IMO) defined these autonomous and unmanned ships as ships that, to a varying degree, can operate independently of human interaction. They are known as Maritime Autonomous Surface Ships (MASS) [2]. The IMO Maritime Safety Committee’s 98th meeting set the degree of autonomy of MASS as follows: A ship at degree one is the same as it is now, the second degree can be operated remotely with the crew on board, the third degree can be operated remotely without the crew on board, and the fourth degree is fully autonomous [3].
Unmanned ships are quite important because they have a variety of commercial [4,5,6,7] and military [8,9] uses, and in particular, considering that most of the causes of ship accidents arise from human errors [10,11], it is very attractive in that it can promote safe navigation [12]. According to Wrobel et al., boarding crews have an advantage in responding to a ship accident, but unmanned ships can significantly lower the likelihood of an accident in advance [13]. Since the ultimate goal of MASS is to be capable of self-decision making, AI-based embedded software is the core of MASS and is expected to replace human seafarers in the future.
Meanwhile, several unique legal issues are expected to emerge as ship operations shift from crew to software in the new era of MASS [14]. Thus far, the negligence of shipowners and crew or ship design and construction defects have been the cause of accidents related to shipping operations. However, once full-scale commercialization of MASS is achieved, defects in autonomous navigation software are likely to be the only factor responsible for accidents. In such cases, the application of existing liability laws and regulations alone would not suffice [5,15,16]. In addition, it is necessary to clarify the legal responsibilities of the stakeholders of autonomous navigation software, considering that a single accident can cause huge losses to the environment, human lives, and property, with multiple stakeholders being involved in the liability and compensation related to such accidents [17,18].
There is a consensus that under the current legal system, to prepare for sustainability-based performance, a marine accident due to a defect in embedded software in autonomous ships will lead to massive legal issues surrounding the burden of proof, the validity of the defenses against liability, and the criteria for defect diagnosis in acknowledging the product liability of the autonomous ship based on the defect in embedded software [19,20]. One can derive important perspectives from many autonomous vehicle research and commercialization cases [21]. Therefore, this study examines the various legal issues arising from defects in the embedded software of autonomous ships based on the UK and Korea product liability laws and explores legislative alternatives as well as methods to improve related laws and legal principles [22,23].

1.2. Research Aim

The aim is to analyze the legal issues in preparation for commercial operations of MASS in the future and to apply comparative methods centered on the Republic of Korea and the United Kingdom. In the case of self-driving cars, a considerable level of self-driving technology has reached the commercialization stage, as shown in Figure 1. In particular, the legal system for self-driving car liability has been prepared to a considerable level based on many studies. Compared to self-driving cars onshore, autonomous ships lack legal schemes and ground despite technological advances that enable coastal navigation and enhance the maritime safety of the maritime industry. Therefore, this study selects five major issues related to product liability, which are the biggest challenges in the commercialization of autonomous ships, and presents solutions to overcome them in Section 4.
Regarding the scope of product liability of embedded software technology, the major legal issues currently pertain to product properties, defect judgment, immunity defense, and burden of proof. Due to the large number of primary and secondary vendors supplying parts and components for MASS to shipyards, it is necessary to establish legal standards and grounds for imposing product responsibility of the centralized integrated algorithm. Therefore, while reviewing the legal theory of the product liability of embedded software, this study compared the legal precedents in the United Kingdom related to the construction of MASS with those in South Korea, a technology leader in marine software and insurance. Based on the analysis, this study compared the legalization of MASS in South Korea and in the UK to recommend the product liability that shipowners, manufacturers, shipyards, class surveyors, insurers, and other stakeholders would bear.

1.3. Current Issues and Research Questions

Legal confirmation of whether product liability is recognized under the UK and Korean laws for various types of embedded software in autonomous ships can serve as a basis for ensuring the commercial operation of autonomous ships in the future [24]. This is because the UK law is most frequently cited as an applicable law for shipbuilding and maritime transport contracts [25]. Additionally, Korea, the world’s leading shipbuilding country that makes huge investments in R&D and commercialization of autonomous ships, cannot be free from product liability. Therefore, when considering the technological factors surrounding autonomous ships, comparative legal investigations and research on Korean and UK laws must be conducted in tandem with technological development.
Intangible assets such as software are manufactured or processed as products (including movable or immovable parts). However, they are not recognized as products under Korea’s Product Liability Act [26]. Nonetheless, with the recent expansion of the role and importance of embedded software in autonomous ships, software should be recognized as a product insofar as it is “embedded” in a specific navigation, engine, or cargo management system. This view is in line with existing research on the recognition of computer software as a product [27].
Against this background, this study conducts a legal analysis and discussion on whether product liability can be applicable in a centrally controlled integrative algorithm by recognizing embedded software on autonomous ships as a system or product. The relevant legal issues/research questions are illustrated in Figure 2, and the legal systems and litigation cases of the Republic of Korea and the UK are compared and analyzed in Section 5 and Section 6.

1.4. Literature Review

Studies on autonomous ships have largely focused on technical perspectives and legal responsibilities.
From the technical perspective, many R&D projects such as MUNIN [28], Revolt [29], Yara Cirkeland [30], and NYK [31] have achieved considerable results. In addition, Lee et al. developed an automatic collision avoidance system and a route-generating algorithm [32], and Song et al. proposed a navigation planning algorithm for unmanned ships in a time-varying environment [33].
From the point of view of accident prevention, there were also several studies to develop accident scenarios and risk assessment methods and response manuals for the failure of autonomous ships, reflecting them in the design, operation, and infrastructure of ships [12,34,35]. According to Ramos et al.’s findings, they identified human errors related to collision avoidance of autonomous ships in remote control using task analysis and tried to reflect them in the design of the autonomous ship [34]. Similarly, Fan et al. proposed a 4p4f (4phase, 4factor) framework to identify the risk influencing factors of MASS operation based on a literature review [12]. Chang et al. identified hazards of MASS operation through the literature review and expert interviews and quantified their scores based on evidential reasoning (ER) and rule-based Bayesian network (RBN) [35]. Fundamentally, these studies believed that human intervention was inevitable even if it was an autonomous ship.
Meanwhile, there are a few studies related to the operational safety of autonomous ships from a legal perspective. For example, Carey cited maritime insurance, COLREG (International Regulations for Preventing Collisions at Sea) compliance, transferring the captain’s role, and pilotage as legal issues for autonomous ships, and reported that unless these problems were solved, the usefulness of autonomous ships would be limited [36]. Ferreira et al. found it necessary to redefine whether an autonomous vessel was a ship [37]. Various studies, including those by Giunta [38], Showalter and Manley [39], Kuwata et al. [40], and Benjamin and Curcio [41], are yet to find a consensus on COLREG compliance due to the ambiguity of the legal status of autonomous ships.
As such, there will be many tasks to be solved in the future for the commercialization of autonomous ships, but research on the product responsibility of autonomous ships is especially urgent and important. This is because the ultimate goal of autonomous ships is an autonomous operation based on the independent judgment of AI software [42]. Depending on how the law is interpreted in practice, the development, application, realization, and commercialization of autonomous technologies will be limited or accelerated.
Delvaux et al. emphasized the importance of product liability in autonomous operations mentioned in the Civil Law Rules on Robotics [43], passed by the European Parliament in 2017. Although the law mentions the need for compensation for damage caused by non-humans or robots, it does not provide detailed guidelines on the identification of liable parties, but only recommends that producers and shipowners prepare for mandatory insurance. Indeed, Leenes et al. found that such measures might not be sufficient [44].
Although some studies have focused on the product liabilities in the maritime sector, most of them were based on the north European Union (EU) legal perspective and there exist limitations to the common views [45]. However, existing research on software manufacturing responsibility for autonomous ships is extremely poor, hence several references in the study of autonomous cars were analyzed. To fill this gap, this study seeks to analyze the legal issues concerning the manufacturing responsibility of embedded software for autonomous ships.
Yeomans indicated that in the UK, primary liability rests with the user of the car, regardless of whether their actions cause the accident or not [46]. If defective technology causes an accident, the user (or their insurer) must pursue the matter legally with the manufacturer. Furthermore, he stated, the current stance could potentially be revised to assign more liability to manufacturers, especially when users were no longer expected to oversee the autonomous driving of their cars.
However, Zipp mentioned that because a fully automated vehicle is expected to make decisions on its own based on its algorithm, manufacturers might argue that it would be impossible to enable such autonomous decision-making capability while still being liable for the vehicle’s accident [47]. He stressed the virtual impossibility of an autonomous vehicle manufacturer being able to conceive of every single situation that might arise when driving.
Kim reviewed the current legal and regulatory framework related to product liability and assessed the challenges in addressing on-board software defects and cybersecurity breaches from both the consumer and manufacturer perspectives [48]. He mentioned that although manufacturers are at the top of assuming responsibility for accidents in fully automated mode, it can be difficult to estimate the scope of liability concerning unexpected software defects; it is also necessary to help consumers address new forms of challenges in product liability claims.
Uzair discussed different liabilities ranging from legal, civil, operator, criminal, moral, product, and insurance to analyze who is liable in case of an automated vehicle accident, as compared to only one or two aspects discussed in the literature [21]. He proposed that liability be attributed so as to benefit all stakeholders as it is a major issue in terms of insurance, and that the manufacturer be held liable for product failure in the absence of other favorable evidence.
Many studies concluded that the manufacturer should, reasonably, assume a part of the liability for embedded software defects in the case of fully autonomous vehicle accidents because as automation technology progresses, these are more likely to be caused by product defects than by driver errors. It is reasonable for manufacturers to assume increased liability for accidents related to mechanical errors.
Thus, based on the above review, this study contributes to filling the gaps in knowledge regarding the product liability of embedded software used in autonomous ships. In addition, it endeavors to make practical contributions to advance the commercialization of future technologies, boost digitization of the maritime industry, and inform about various social and legal benefits for all stakeholders.

2. Research Theory and Methods

2.1. Theory

From a time-series standpoint, the product liability law came about in the 20th century in recognition of the risk and safety of product use, general knowledge and anticipation of consumer risk, and strict liability through warning indication. Therefore, it aimed to directly transfer, from the manufacturer to the consumer, the product liability for damages caused by product defects in accordance with the privacy of the product contract. With the formulation and enactment of the Restatement of the Law Second, Torts, by the American Law Institute in 1965, product liability has not only placed on manufacturers the burden of the liability for damages but also influenced legislation based on no-fault liability in advanced countries such as the UK, the EU, Japan, and Korea.
The purpose of interpreting and applying the comparative law theory to the legal scheme of the Republic of Korea and the UK in relation to the product liability law is to present an alternative to abstract research in the field of maritime law, which greatly needs legal stability. As a legal principle, a product liability law for MASS is set to play an important part in the digital shipping revolution 4.0, where information on shipbuilding technology is currently asymmetric between shipyards and shipowners. Therefore, by introducing a conceptual definition of autonomous ships and confirming decisive similarities and differences between countries through a comparative study, this paper makes a meaningful contribution [49].

2.1.1. Korea’s Product Liability Act and Its Application to Autonomous Ships

Product liability refers to the liability for damages borne by a consumer/user or any person involved in the manufacture and sale of the product, such as a manufacturer, vendors, distributors, suppliers, retailers, and other stakeholders who make products available to the public and are held responsible for damages incurred to the life, body, or property of a third party due to defects in the product. This legal principle of product liability was established based on the principle of strict liability formed by precedents in the United States in 1960. After approximately 25 years of debate and delay, the European Council Directive adopted 85/374/EEC in 1985, Japan enacted the Product Liability Act in 1994, and many developed countries followed suit, in an attempt to specify the responsibility of the modern industrial society to replace and supplement classical tort liability and default liability, as shown in Table 1.
At present, it is premised that the product defect is linked to the manufacturer’s direct responsibility for the consumer’s consequential damage. Korea’s Product Liability Act stipulates responsibility by guaranteeing freedom of individuals or companies and complementing or replacing existing laws in consideration of technological changes and diversification of responsible subjects. According to Article 2(1) of the Act, products are defined as manufactured or processed movable assets, including those that constitute other movable assets or parts of real estate. If damage occurs due to a defect in the MASS, it will be treated under movable property in the design and construction of the ship and included in the scope of product liability. In addition, according to Article 9(1) of the Standard Shipbuilding Agreement of the Shipbuilders’ Association of Japan (SAJ), commonly used by domestic shipyards: “The shipyards shall repair defects in ships found within 12 months of ship delivery at their own expense”.
In particular, since a MASS is assembled with many parts, even in the case of damage caused by defects in specific parts constituting the movable property, the scope of responsibility is inevitably expanded top-down to the shipyard, the manufacturer, or the manufacturer of the part. Chapter 4 reviews whether it is possible to expand product liability to defects in embedded software or communication errors between autonomous ships and land.

2.1.2. The UK Product Liability Act and Its Application to Autonomous Ships

The Consumer Protection Act was adopted by the UK on 15 May 1987, after the European Council (EC) Directive 85/374 was enacted and took effect on 1 March 1988. In Chapter I of the Consumer Protection Act 1987, the UK stipulated the rights and obligations of consumers affected by defects in products. The UK has revised and expanded the Consumer Protection Act 1987 (Modification) on 4 December 2000, by recognizing product liability for agricultural products in accordance with the EC legislative guidelines revised on 10 May 1999. In particular, the UK Product Liability Act accepts the EC legislation guidelines by stipulating that Chapter I of the UK Consumer Protection (Part I) can be interpreted in compliance with the EC legislation guidelines. In addition, according to Chapter 1 of the UK Consumer Protection Act (Part I), a product refers to movable property, electricity, or a product that is comprised of another product in the form of components, raw materials, or otherwise. According to the current UK Consumer Protection Act, product liability cannot be applied to the software design itself. However, in the case of embedded software installed in navigation, a cargo control system is essential for the operation of MASS, so that the product liability for MASS can be clearly established.

2.2. Dual Comparison Method

In the process of legal comparison, it is useful to clarify the method of “vertical” comparative law being adopted as different methods should consider varied risk factors [50]. Vertical legal comparison is defined as the transposition of legal concepts, or the ideas behind them, from the national to the international level. This study proposes that legal comparisons for the product liability of MASS would be primarily “vertical” as it is beneficial to analyze national legal systems to enhance the understanding of international stakeholders. Figure 3 shows a concept of the research method. The vertical method for a comparative law analysis of national systems is justified by the emergence of a “common zone” of product liability in MASS. Finally, this study suggests a guide when designing and implementing vertical comparative law methods for all stakeholders.

3. Results

This section reviews the comparative findings of Korean and UK product liability laws applicable to the embedded software in MASS.

3.1. Korea

3.1.1. Subjects and Scope

According to Article 2(1) of the Product Liability Act, the term “product” refers to manufactured or processed movables (including those forming part of other movables or immovables). Movables are any objects except for immovable properties, materials of a shape in a solid, liquid, or gaseous state, including intangible energy such as electricity and heat. Movables can be either a finished product or any parts thereof, or its raw material and include new products, used and recycled products, mass-produced industrial goods, handicrafts, and artworks.
In this context, embedded software units installed on autonomous ships are the result of a manufacturing process involving a series of actions, including product design, manufacturing, inspection, and labeling. Interpreting the definition in Article 2(1), properties are immovable products (real estate) such as apartments, buildings, and bridges, but facilities for lighting, plumbing, and air conditioning, as well as elevators, windows, and the like fall under the category of movables.
According to Article 2(3) of the Product Liability Act, the term “manufacturer” refers to either (a) a person engaged in manufacturing, processing, or importing products or (b) a person indicated as the person under (a) by labeling the products with a personal name, firm name, trademark, or any other sign, including misleading indications, and even a seller of products whose manufacturer the victims cannot identify.
Consequently, pursuant to Article 3(1) of the Act, a person who bears responsibility is liable for compensation for the damages caused by a defect in a product. Pursuant to Para 2 of the same article, a manufacturer has the obligation to take necessary measures against a defect in a product. Article 4 of the Act provides eligible defenses against liability: (1) the manufacturer did not supply the product, (2) the existence of the defect could not be identified by the scientific or technical knowledge at the time of product supply, (3) the defective product was manufactured in compliance with the statutory standard of the time, and (4) in the case of raw materials or components, the defect was caused by the design or manufacturing instruction by the manufacturer of the relevant raw materials or components.

3.1.2. Korean Courts’ Position on Product Liability

Under the Product Liability Act, the compensation paid as damages to consumers due to malicious, illegal acts has been rather low so far, compared to the profits made by manufacturers because of the difficulty for the victims to present scientific or technical proof of the defect. However, according to the Korean Supreme Court’s decision on 25 February 2000, (98-da-15934), the Product Liability Act serves the purpose of a fair and reasonable compensation system to ease the burden of proof of the consumer based on the presumption that if an accident occurred while the product was being used normally, the accident occurred due to a defect in the product. As shown in Table 2, this ruling suggests that the consumer’s burden of proof is reduced by proving that the damage occurred during the normal use of the product, which allowed the assumption that the product had a defect at the time of project delivery.
With the amendment to the Product Liability Act on 22 May 2013, a system of punitive damages was introduced under Article 3(2), wherein a supplier could be held liable for damage arising from product defects under Article 3(3). Thus, not only did the courts concretize the product liability of the manufacturer, but also expanded the liability of the supplier when the manufacturer cannot be identified. Additionally, with the insertion of Article 3–2 (Presumption of Defects) on 18 April 2017, the courts eased the consumer’s burden of proof by legislating the presumption of defects in the causal relationship between product defects and damage. However, despite the newly introduced legal provisions on the presumption of defects, the courts still do not assign unconditional liability to the manufacturer for the damage caused by the product. The manufacturer is held liable only when the product has a defect, and the damage is caused by that defect.
In addition, regarding the recognition of software as a product in Korea, in principle, software is immaterial and thus not considered a movable thing but a service. There are two opposing views on software defects. According to the negative view, a software defect can be resolved by clarifying the damage arising from negligence as tortious conduct under the Civil Act. The opposing and affirmative view argues that software is a product in the broad sense of the word and according to the definition of “things” in Article 98 of the Civil Act, which also recognizes electricity as a thing.
Future controversy surrounding embedded software used in autonomous ships is expected in Korea, given the lack of a clear legal framework and precedents under the Product Liability Act for defects in the AI-based embedded software system of autonomous ships.

3.2. The United Kingdom

3.2.1. Target and Scope

Section 1(2)(c) of the UK Consumer Protection Act 1987 with the section heading “Purpose and construction of Part I” stipulates that “product” refers to goods or electricity and includes a product-forming part of another product as a component or raw material or otherwise. Regarding the definition of “defect”, Section 3(1) with the section heading “Meaning of defect” reads: “(1) Subject to the following provisions of this section, there is a defect in a product for the purposes of this Part if the safety of the product is not such as persons generally are entitled to expect; and for those purposes, “safety” in relation to a product shall include safety with respect to products comprised in that product and safety in the context of risks of damage to property, as well as, in the context of risks of death or personal injury. (2) In determining for the purposes of subsection (1) above, what persons generally are entitled to expect in relation to a product, all the circumstances shall be taken into account, including: (a) the manner in which and purposes for which the product has been marketed, its get-up, the use of any mark in relation to the product and any instructions for, or warnings with respect to, doing or refraining from doing anything with or in relation to the product; (b) what might reasonably be expected to be done with or in relation to the product; and (c) time when the product was supplied by its producer to another; and nothing in this section shall require a defect to be inferred from the fact that the safety of a product that is supplied after that time is greater than the safety of the product in question”.
Accordingly, the UK Consumer Protection Act 1987 shares the same legislative purpose as the manufacturer’s “liability without fault” stipulated by the 1999 EC Directive (EC Directive 99/34) for liability for defective products with a more detailed enumeration of the subjects and scope.

3.2.2. Position of UK Courts on Product Liability

In the UK, there are hardly any product liability litigations associated with software defects. Therefore, this subsection first examines the ruling on Hamble Fisheries Ltd. v L Gardner and Sons Ltd. (The “Rebecca Elaine”) (1999) case 2 Lloyd’s Rep. 1, a case dealing with the duty of warning about product defects.
In this case, a fishing company (plaintiff), which possesses various types of fishing vessels, entered into a contract with a shipbuilder to build a fishing vessel, the Rebecca Elaine. The vessel was fitted using a Gardener engine. However, the Gardener engine manufacturer entered into a sales contract with the shipbuilder in 1985 instead of directly selling it to the plaintiff company. The vessel owner took over the new ship fitted with the Gardener engine with a manufacturer’s warranty of one year. A year after selling the vessel, the defendants became aware of the malfunctioning piston of the Gardener engine after receiving reports that the engine was breaking down. However, the defendants did not warn the plaintiff, and Rebecca Elaine’s engine failed well before the guaranteed time of 20,000 h. The Gardener engine was manufactured by Wellworthy Ltd., taken over by L Gardner and Sons Ltd. (defendants).
Justice Nourse, Lord Justice Mummery, and Lord Justice Tuckey of the Court of Appeal (Lord Justice) emphasized that the piston broke down even before the lapse of 1/3 of the guaranteed operating time specified in the piston manual, which said that “the engine’s piston would run for 20,000 h or more without needing to be dismantled or replaced”. The court held that when the defendants took over Wellworthy, they were clearly informed that the defective piston could cause significant engine failures.
However, the defendants ignored the warnings and did not take any measures to warn the vessel owner about the engine’s condition. In fact, the Rebecca Elaine was emergency towed 4 miles southwest of Needles, England, after the engine stopped at sea. The court paid attention to the fact that the defendant—the manufacturer of the Gardener engine—did not warn the plaintiff against the latent defect of the main engine, the core equipment for the new vessel, and held the defendants liable for negligence and damage sustained by the plaintiff. Conclusively, the UK Court of Appeals held that manufacturers have a duty to warn consumers about safety defects.
The significance of this court decision lies in that it established that manufacturers owe the duty of warning consumers to follow instructions for safety defects by virtue of the stipulation in Section 3 “Meaning of defect” of the Consumer Protection Act 1987, which states that a predictable latent defect of a product includes product safety against the product-related safety risk and property loss and in light of the criteria for determining what is generally expected of the product in question, such as the methods of and purposes for the product marketing, product composition, use of marks and the instructions or warnings with respect to doing or refraining from doing anything in relation to the product.

4. Potential Issues in Embedded Software Defects in the Age of Autonomous Ships

4.1. Advancement to RQ No. 1

In the future, autonomous ships are expected to configure decision making on collision avoidance, route selection for navigation, typhoon avoidance, vessel speed adjustment, cargo management, and vessel security by remotely receiving information from the land with the support of the e-navigation system and by interpreting the data on the surrounding environment collected by built-in sensors using an AI-based centrally controlled algorithm [51].
In optimizing the technology for autonomous ship operation, navigation and engine equipment with built-in software cannot fulfill its functions properly only with the mechanical durability of hardware, which is an external requirement. Only when the autonomous ship is connected to its embedded software, can the same level of seaworthiness as that of a vessel with crew on board be ensured. Based on this premise, embedded software can be integrated into hardware such as hulls and other equipment and can thus be recognized as a movable property, which is covered under the Product Liability Act. Embedded software on autonomous ships is not a generic package software, but software installed in the product itself, which makes it impossible for the software to perform its functions independently. Rather, embedded software performs its functions such as route setting, course trajectory adjustment in relation to other ships, typhoon avoidance, and cargo management only when it is embedded in the equipment units of an autonomous ship. This proves its inseparability from the vessel equipment units in which it is embedded.
When the shipowner provides owner-furnished equipment with embedded software to the shipyard that has placed an order, the software and the ship equipment units are not recognized as separate objects of transaction but as integral parts of the objects essential for performing the functions necessary for the operation of an autonomous ship. In this context, the embedded system, which combines embedded software and vessel equipment, is considered a movable property, to which the legal status of a property is assigned under the Product Liability Act based on circular reasoning.
The application of product liability to general software can be established in accordance with Section 2 “Liability for defective products” and Paras (2)–(5) of the UK Consumer Protection Act 1987 that states: “Where two or more persons are liable by virtue of this Part for the same damage, their liability shall be joint and several” Section 2(5). By virtue of this “joint and several” liability clause, if software embedded in an equipment unit operated by the centrally controlled algorithm, such as that of an autonomous ship, is recognized as an integral component of another equipment unit, the application of the product liability for the embedded software can be expanded by virtue of the product liability regulations based on the manufacturer’s tortious conduct.
As there is much room for varied interpretations of the concept of seaworthiness of automated ships judged from the perspective of marine insurance, the requirements for determining the level of safety to be generally expected from a ship as a manufactured product need to be reviewed more closely.

4.2. Advancement to RQ No. 2

By recognizing a fault or defect in embedded software as “material”, the pattern of loss transfer can be maintained according to the principle of “liability without fault”, by which the victim claims damages for the use of the defective product under the Product Liability Act. However, in the case of embedded software, to assume product liability in a state separate from the product in which it is embedded, it must be evaluated or proven by a third-party verification agency that the embedded software is independent of the product in which it is embedded [52].
In reality, however, it is a highly challenging process to uniformly and clearly specify the standards for evaluating the independence between embedded software and the product concerned in terms of laws and regulations. Nevertheless, in case of a dispute over this classification, it is necessary to consider several criteria to assess the independence between embedded software and the autonomous ship. First, the embedded software manufacturer and the manufacturer of the product in which it is embedded should be separate entities. For example, if an autonomous ship is built by the shipyard “A”, the LiDAR sensor installed on the ship should be manufactured and supplied by company “B” under a separate contract. This is because if the supplier and manufacturer are recognized as the same entity, the embedded software is included in the product liability with no room for dispute. Second, the embedded software should be able to upgrade or update the firmware without changing the tangible storage medium. For example, in the case of a Level 4 unmanned autonomous ship, which is navigable by AI-based decision making alone, periodic updates or upgrades of embedded software and protection work such as anti-virus inspection, should be performed independently. In this process, the embedded software must be separated from the hardware, such as specific navigation communication equipment, engine equipment, and cargo management equipment. Third, for the periodic update of the embedded software, as mentioned in the second criterion, there must exist an input/output interface through which the embedded software can be accessed. If the interface is not accessible, then the embedded software is completely incorporated into the product, which leaves no option but to recognize it as an integral part of the product.
In conclusion, it is impossible to separate the embedded software from the product and limit the product liability, as mentioned in Section 4.1, based on the legal definition of software. For example, just as the mast and hull of a sailing ship are interconnected as compound and appurtenance, the autonomous ship is comparable to the compound and embedded software to the appurtenance.

4.3. Advancement to RQ No. 3

In this subsection, a legal review is performed to determine whether software pertains to the product in which it is embedded under the Product Liability Act and, if it does, who assumes the legal responsibility when the user or a third party sustains extended damage due to a fault or defect in the embedded software. As discussed in Section 4.1 and Section 4.2, embedded software does not have an independent function in itself but performs the dedicated functions only when it is embedded in a specific predetermined product or hardware. In this sense, it is desirable to recognize the product property of embedded software by considering the embedded software as an independent movable property.
In the future, for damages caused by a fault or defect in software embedded in products based on multiple state-of-the-art technologies such as autonomous ships, shipowners will be expected to prove the intentional torts and negligence to the manufacturer in real time with respect to the defects during operation. In addition, the shipowner must make up for the loss by structuralizing the process of claiming damages from the product manufacturer by virtue of “liability without fault”, the core principle of product liability.
Eventually, the problem of defects in embedded software should be resolved in the form of compensation for damages by default or liability for defects based on the internal subcontracting arrangement between the shipyard, the manufacturer of autonomous ships, tangible material, and the embedded software equipment company.

4.4. Advancement to RQ No. 4

Article 60 (Defect Warranty Obligations for Software Projects) of the Software Promotion Act, which is a prime example of terrestrial cases, stipulates that a software business entity, which has entered into a software project contract with a state agency, shall bear warranty obligations for defects that occur within one year from the projection completion date (i.e., the date of delivery of the final software output upon completion of tests and inspections of the project), with a proviso that if the state agency has separately placed an order for a software project under Article 44(3), the contracting parties may otherwise determine the defect warranty obligations.
The duration of warranty obligations set at one year, compared to two years for hardware or general projects, is opposed as being against equity among business operators. Furthermore, in Article 7(1) of the Product Liability Act of Korea, the right to claim damages expires three years after the date on which the victim becomes aware of the damage. For example, considering the legislative purpose of the Product Liability Act, if the governing law for contracts and insurance related to autonomous ships is designated as that in Korea, the shipowner and victim should check whether the period for warranty obligations is included in the shipbuilding contract. In addition, from the perspective of the shipowner, it is desirable to insert a clause regarding the effective period of warranty based on a separate agreement with the shipyard in preparation for the risk from defects in the embedded software system of the autonomous ship.
Shipbuilding contracts such as the SAJ Form, the AWES (Association of European Shipbuilders and Shiprepairers) Form, BIMCO (Baltic and International Maritime Council), and NEWBUILDCON 2008 (standard newbuilding contract), all typical examples of maritime contracts, stipulate that the shipyard should repair every fault or defect within one year after delivery to the shipowner, set as the period of warranty for new ships. However, whether this clause should be applied to design software defects is still a topic for discussion, given that most of the physical defects are associated with construction defects, and, as seen in the UK case of Aktiebolaget Gotaverken v Westminster Corporation of Monrovia and Anor (1971) 2 Lloyd’s Rep 505, Donaldson J. of Queen’s Bench Division (Commercial Court), which “…required good workmanship both in the design and the execution” and extended the scope of defects from physical to non-physical elements. Swedish Shipbuilders’ Association General Regulations (1956) no. 11 and 13 relate to whether the ship-repairer is liable for the detention of the vessel that is under repair in the shipyard. The Commercial Court considered that the repair shipyard had a warranty of obligations in respect of defects or deficiencies of material, workmanship, and shipyard design errors; any shipbuilding design error should be held accountable as a manufacturing problem, and, therefore, there was no reason for it to be covered by the warranty clause. Notwithstanding the Commercial Court judgment, the parties provided expressly that the builder’s warranty covers defects resulting from design faults in the duration of warranty obligations.

4.5. Advancement to RQ No. 5

One of the primary legislative purposes of Korea’s Product Liability Act is to specify the grounds for exemption from liability for fair distribution of risks between the victim and the manufacturer. Article 4 (Exemptions) of the Product Liability Act stipulates four defenses: non-supply of product, development risk, compliance with laws and regulations, and design and manufacturing instruction by the manufacturer. Therefore, if Korean law is designated as the law governing a case of product liability for embedded software, which is the focus of this study, the shipyard (manufacturer) is exempted from product liability if it can be proved that the scientific and technical state at the time of delivery of the autonomous ship did not allow them to detect the defect during the construction. This exemption clause has a significant implication in that it provides opportunities for manufacturers to defend themselves in the face of the risk to the development of new technologies, with latent risks associated with autonomous vessels. This defense against liability has both advantages and disadvantages. Manufacturers can perform R&D and the implementation of autonomous ships more actively supported by this exemption clause; however, the purpose of consumer protection, the raison d’être of the Product Liability Act, could be compromised in the process.
In other words, for an AI-based autonomous ship with no crew on board that has not yet begun commercial operation, it is necessary to legally review the grounds for exemption by striking a balance between the two conflicting stakeholders: manufacturers and consumers. In addition, with respect to the burden of proof, although the Korean Product Liability Act recognizes the principle of “liability without fault”, there are no clear provisions governing the burden of proof. Consequently, the victim must prove: the product defect, the product defect causing damage, and a causal relationship between the defect and the damage. In the case of the embedded software used in autonomous ships, which is extremely technology-intensive with the design and manufacturing process subject to patents and personal information protection, it is even more difficult to prove whether material defects lead to expanded damage, especially in circumstances where victims find themselves in an inferior position than manufacturers in social and economic terms. As small and medium-sized shipowners, victims lack specialized knowledge compared to shipyards, which are the manufacturers.
As defenses against product liability for UK manufacturers, Article 4(1) of the Consumer Protection Act 1987 provides several options: (1) detection is attributable to compliance with legal requirements; (2) the manufacturer did not supply the product at any time to another consumer; (3) the only supply of the product was to another by the manufacturer otherwise than in the course of a business of that person and the manufacturer does not apply to that person or applies to him by virtue only of things done other than with a view to profit; (4) the defect did not exist in the product at the relevant time; (5) the state of scientific and technical knowledge at the relevant time was not such that a producer of products of the same description as the product in question might be expected to have discovered the defect if it had existed in his products while they were under his control; and (6) the defect constituted a defect in a product as its constituent and was wholly attributable to the design or compliance of its producer of the former product with instructions given by the producer of the latter product.
The law is structured in such a manner that the proof of any one of the above options is sufficient to exempt the manufacturer from product liability. Similar to Korea, the defenses granted by the UK Product Liability Act, which integrated the European Council Directive 85/374/EEC on liability for defective products into domestic legislation, proves the existence of a reasonable risk by setting the scientific and technical state of the producer of the product in question [52].

5. Discussion

Our main findings regarding legal and policy implications were derived by examining relevant Korean and UK laws, and case laws on software-related product liability. Based on the level of autonomy of MASS, we found that the main body of responsibility lies with the shipowners from the perspective of navigation assistance for Levels 1 and 2. When an embedded software system impacts the MASS, the main body of responsibility lies with the manufacturer or vendor. Based on the results, this study presented the criteria for product liability recognition and desirable legal and policy-related improvement measures, focusing on (i) recognition of software as a product, (ii) defect assessment criteria for (i), and (iii) recognition of technical risk defense as a defense against software product liability.
A comparative legal analysis revealed that there are hardly any legislation or litigation cases explicitly applying the product liability law or system. Nevertheless, all countries share awareness of the problem that embedded software in MASS can cause unexpected risks of life, body, and property, and that it is not appropriate to resolve the problem through contract law between vendors, shipbuilders, owners, and shipyards, as shown in Figure 4.
As part of this institutional review, it is important to understand the Product Liability Act from different perspectives. The primary issue is that product liability is discussed in relation to all software types distributed online over wired or wireless networks as movable properties. This includes even the parts not directly connected to the media. However, even though software supplied as a medium is understood and treated as a product subject to product liability if it is supplied online, it is interpreted as a service contract, to which tortious negligence is applied, imposing the burden of proving the supplier’s negligence, inevitably criticized as a biased interpretation. Taking into account this criticism, it is recommended that software, such as electricity, be included in the category of “product”.
However, regarding material damage related to software embedded in autonomous ships, in-depth research needs to be conducted to understand whether shipowners can claim damages from software manufacturers for negligence, leveraging their direct right of action in preparation for future occurrences of physical damages in relation to embedded software at the level of marine insurance and shipbuilding insurance.
As a limitation of this study, it may be emphasized that in case of an accident caused by a defect in software embedded in an autonomous ship, even if there is no problem in recognizing product liability in the embedded software as a defect in the autonomous system, the embedded software itself is treated as a part or component, which entails the question on imposing product liability separately on centrally controlled integrative algorithm developers or companies. Therefore, it is necessary to analyze the possibility of proportional application of legal principles to cases where the shipowner, who manages an autonomous ship owned by company “A” and remotely operated by centrally controlled integrative algorithms delivered by companies “B”, “C”, and “D”, can select any of these algorithms for a safe future for the maritime industry.

6. Conclusions

Today, with rapid technological development, a huge paradigm shift has occurred in which technologies once understood as concepts of information are gradually overtaking our lives, making it necessary to discuss the legal context of the universality of information, while breaking away from the limited perception of software in the maritime industry.
For this purpose, legal and policy alternatives for ensuring maritime safety of the shipping and shipbuilding countries need to be analyzed and modified. In examining the interpretation approaches and legislative theories regarding the product liability law in both Korea and the UK, although it was not possible to derive specific legal and policy implications, this study identified various legal issues related to software product liability and macroscopic legal policy directions. Based on these findings, the following legal interpretations and policy conclusions can be drawn, along with legislative policy directions. First, as a procedure for determining software product liability, we presented stepwise assessment process criteria for each step to determine whether the software in question (i) was manufactured, (ii) was a movable thing, (iii) was recognized as an independent entity, especially as an embedded type, by preserving its unique properties, (iv) had the possibility of reasonable expectation of an alternative design in the software, (v) was embedded in a defective storage medium, (vi) was properly labeled with risks, (vii) involved a third party other than its manufacturer, and (viii) complied with the reliability standard protocol.
Furthermore, we made the following proposals and presented concrete implementation strategies: (i) inserting new software-related provisions into the Product Liability Act as a policy alternative to resolve the controversy surrounding the physical properties of software in relation to product liability recognition, (ii) establishing regulations to reasonably limit the scope of software product liability in the Software Promotion Act, (iii) introducing an insurance system and a fund to expand the scope of application of software product liability, and (iv) preparing an alternative dispute resolution procedure with expert participation to achieve ideological harmony and balance between equitable damage relief and advancement of related technologies.
With the rapid development of AI and the increase in the independence of the software itself, the damage to life, body, and property associated with software defects is also expected to increase. To efficiently cope with such situational and technological changes and contribute to the safety of the maritime industry, we recommend that the legal and policy alternatives proposed in this study be implemented in the legal system. This will lead to appropriate safeguarding of the institutional ideal of product liability in the area of software and consumer rights while welcoming technological progress by reducing its impact on promoting inclusive sustainable growth and development. In addition, since ships operate internationally, a variety of stakeholders are intertwined. From this point of view, this study examined the potential liability issues caused by defects in the AI software embedded in MASS and draw implications for solving them. These results are expected to contribute to future amendments of international conventions and marine insurance agreements.

Author Contributions

Conceptualization, D.K. and S.P.; Methodology, C.L. and S.L.; Validation, D.K. and C.L.; Investigation, S.L.; Resources, S.L.; Data Curation, D.K.; Writing Preparation, S.L. and S.P.; Writing—Review and Editing, D.K. and S.L.; Visualization, C.L.; Supervision, S.P.; Project Administration, S.L.; Funding Acquisition, D.K. All authors have read and agreed to the published version of the manuscript.

Funding

This research was a part of a project titled “Development of Smart Port-Autonomous Ships Linkage Technology [grant number 1525012737]”, funded by the Ministry of Oceans and Fisheries, Korea.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Petrig, A. Unmanned Offender and Enforcer Vessels and the Multi-Dimensional Concept of ‘Ship’ under the United Nations Convention on the Law of the Sea. In UNCLOS as a System of Regulation, 1st ed.; Kristina, S., Fenella, B., Birgit, F., Eds.; Routledge: London, UK, 2021. [Google Scholar]
  2. IMO Takes First Steps to Address Autonomous Ship. Available online: https://www.imo.org/en/MediaCentre/PressBriefings/Pages/08-MSC-99-MASS-scoping.aspx (accessed on 15 January 2022).
  3. Autonomous Ships: Regulatory Scoping Exercise Completed. Available online: https://www.imo.org/en/MediaCentre/PressBriefings/pages/MASSRSE2021.aspx (accessed on 15 January 2022).
  4. Porathe, T. A Navigating navigator onboard or a monitoring operator ashore? Towards safe, effective, and sustainable maritime transportation: Findings from five recent EU projects. Transport. Res. Proc. 2016, 14, 233–242. [Google Scholar] [CrossRef] [Green Version]
  5. Kretschmann, L.; Burmeister, H.C.; Jahn, C. Analyzing the economic benefit of unmanned autonomous ships: An exploratory cost-comparison between an autonomous and a conventional bulk carrier. Res. Transp. Bus. Manag. 2017, 25, 76–86. [Google Scholar] [CrossRef]
  6. Ramos, M.A.; Utne, I.B.; Vinnem, J.E.; Mosleh, A. Accounting for Human Failure in Autonomous Ship Operations. In Safety and Reliability—Safe Societies in a Changing World, 1st ed.; Haugen, S., Barros, A., van Gulijk, C., Kongsvik, T., Vinnem, J.E., Eds.; CRC Press: London, UK, 2018. [Google Scholar] [CrossRef]
  7. Rodseth, O.J. From concept to reality: Unmanned merchant ship research in Norway. In Proceedings of the Underwater Technology (UT), Busan, Korea, 21–24 February 2017; p. 10. [Google Scholar] [CrossRef]
  8. Chang, Y.; Zhang, C.; Wang, N. The international legal status of the unmanned maritime vehicles. Marine Policy 2020, 113, 103830. [Google Scholar] [CrossRef]
  9. Schmitt, M.N.; Goddard, D.S. International law and the military use of unmanned maritime systems. Int. Rev. Red Cross 2016, 98, 567–592. [Google Scholar] [CrossRef]
  10. Rothblum, A.M. Human Error and Marine Safety, National Safety Council Congress and Expo, Orlando. 2000. Available online: http://www.bowles-langley.com/wp-content/files_mf/humanerrorandmarinesafety.pdf (accessed on 8 March 2021).
  11. Utne, I.B.; Sorensen, A.J.; Schjolberg, I. Risk management of autonomous marine systems and operations. In Proceedings of the International Conference on Offshore Mechanics and Arctic Engineering, Hamburg, Germany, 5–10 June 2017; pp. 1–10. [Google Scholar] [CrossRef]
  12. Fan, C.; Wróbel, K.; Montewka, J.; Gil, M.; Wan, C.; Zhang, D. A framework to identify factors influencing navigational risk for Maritime Autonomous Surface Ships. Ocean. Eng. 2020, 202, 107188. [Google Scholar] [CrossRef]
  13. Wrobel, K.; Montewka, J.; Kujala, P. Towards the assessment of potential impact of unmanned vessels on maritime transportation safety. Reliab. Eng. Syst. Saf. 2017, 165, 155–169. [Google Scholar] [CrossRef]
  14. Ringbom, H.; Rosaeg, E.; Solvang, T. Autonomous Ships and the Law, 1st ed.; Routledge: London, UK, 2020. [Google Scholar]
  15. Coito, J. Maritime Autonomous Surface Ships: New Possibilities—and Challenges—in Ocean Law and Policy. Int. Law Stud. 2021, 97, 259–306. [Google Scholar]
  16. Ghaderi, H. Autonomous technologies in short sea shipping: Trends, feasibility and implications. Transp. Rev. 2018, 39, 152–173. [Google Scholar] [CrossRef]
  17. Caccia, M. Autonomous surface craft: Prototypes and basic research issues. In Proceedings of the 14th Mediterranean Conference on Control and Automation, Ancona, Italy, 28–30 June 2006; pp. 1–6. [Google Scholar] [CrossRef] [Green Version]
  18. Liu, Z.; Zhang, Y.; Yu, X.; Yuan, C. Unmanned surface vehicles: An overview of developments and challenges. Annu. Rev. Control 2016, 41, 71–93. [Google Scholar] [CrossRef]
  19. Comite Maritime Internation Position Paper on Unmanned Ships. Available online: https://comitemaritime.org/wp-content/uploads/2018/05/CMI-Position-Paper-on-Unmanned-Ships.pdf (accessed on 1 November 2021).
  20. Ringbom, H.M.; Veal, R. Unmanned ships and the international regulatory framework. J. Int. Marit. Law 2017, 23, 100–118. [Google Scholar]
  21. Uzair, M. Who is liable when a driverless car crashes? World Electr. Veh. J. 2021, 12, 62. [Google Scholar] [CrossRef]
  22. Li, S.; Fung, K.S. Maritime Autonomous Surface Ships (MASS): Implementation and legal issues. Marit. Bus. Rev. 2019, 4, 330–339. [Google Scholar] [CrossRef] [Green Version]
  23. Vojkovic, G.; Milenkovic, M. Autonomous Ships and Legal Authorities of the Ship Master. Case Stud. Transp. Policy 2020, 8, 333–340. [Google Scholar] [CrossRef]
  24. Nava, G.; Fernanda, P.M. Legal Challenges of Liability in Collisions Arising from the Development of Autonomous and Unmanned Shipping—International and Norwegian Perspective. Master’s Thesis, University of Oslo, Oslo, Norway, 2019. [Google Scholar]
  25. Chamberlain, A.; Colaco, H. The Shipping Law Review: United Kingdom—England & Wales. Law Revs 8 June 2021. Available online: https://thelawreviews.co.uk/title/the-shipping-law-review/united-kingdom-england--wales (accessed on 1 November 2021).
  26. Jae-Seon, S.; Jung-Eun, S. Product liability in Korea. In Product Liability: Fundamental Questions in a Comparative Perspective; Koziol, H., Green, M.D., Lunney, M., Oliphant, K., Yang, L., Eds.; De Gruyter: Berlin, Gemany, 2017; pp. 80–96. [Google Scholar] [CrossRef]
  27. Beard, T.R.; Ford, G.S.; Koutsky, T.M.; Spiwak, L.J. Tort liability for software developers: A law & economics perspective. J. Marshall J. Comput. Inf. Law 2009, 27, 199. [Google Scholar]
  28. MUNIN. Research in Maritime Autonomous Systems Project Results and Technology Potentials. 2016. Available online: http://www.unmanned-ship.org/munin/wp-content/uploads/2016/02/MUNIN-final-brochure.pdf (accessed on 20 January 2022).
  29. Alfheim, H.; Muggerud, K. Development of a Dynamic Positioning System for the ReVolt Model Ship. Master Thesis, Norwegian University of Science and Technology, Trondheim, Norway, June 2017. Available online: https://ntnuopen.ntnu.no/ntnu-xmlui/bitstream/handle/11250/2452115/16486_FULLTEXT.pdf (accessed on 20 January 2022).
  30. Kongsberg, Autonomous Ship Project, Key Facts about YARA Birkeland. 2017. Available online: https://www.kongsberg.com/maritime/support/themes/autonomous-ship-project-key-facts-about-yara-birkeland (accessed on 11 March 2022).
  31. NYK Line. NYK Conducts World’s First Maritime Autonomous Surface Ships Trial. 2019. Available online: https://www.nyk.com/english/news/2019/20190930_01.Html (accessed on 11 March 2022).
  32. Lee, M.C.; Nieh, C.Y.; Kuo, H.C.; Huang, J.C. An automatic collision avoidance and route generating algorithm for ships based on field model. J. Mar. Sci. Technol. 2019, 27, 101–113. [Google Scholar] [CrossRef]
  33. Song, R.; Liu, Y.; Bucknall, R. A multi-layered fast marching method for unmanned surface vehicle path planning in a time-variant maritime environment. Ocean Eng. 2017, 129, 301–317. [Google Scholar] [CrossRef]
  34. Ramos, M.A.; Utne, I.B.; Mosleh, A. Collision avoidance on maritime autonomous surface ships: Operators’ tasks and human failure events. Saf. Sci. 2019, 116, 33–44. [Google Scholar] [CrossRef]
  35. Chang, C.H.; Kontovas, C.; Yu, Q.; Yang, Z. Risk assessment of the operations of maritime autonomous surface ships. Reliab. Eng. Syst. Saf. 2021, 207, 107324. [Google Scholar] [CrossRef]
  36. Carey, L. All hands off deck? The legal barriers to autonomous ships. In NUS Centre for Maritime Law Working Paper 17/06; Centre for Maritime Law (CML): Singapore, 2017; p. 8. [Google Scholar] [CrossRef]
  37. Ferreira, F.; Alves, J.; Leporati, C.; Bertolini, A.; Bargelli, E. Current regulatory issues in the usage of autonomous surface vehicles. In Proceedings of the OCEANS—MTS/IEEE Kobe Techno-Oceans (OTO), Kobe, Japan, 28–31 May 2018; pp. 1–9. [Google Scholar] [CrossRef]
  38. Giunta, L. The enigmatic juridical regime of unmanned maritime systems. In OCEANS 2015-Genova; IEEE: New York, NY, USA, 2015; pp. 1–9. [Google Scholar] [CrossRef]
  39. Showalter, S.; Manley, J. Legal and engineering challenges to widespread adoption of unmanned maritime vehicles. In OCEANS; IEEE: New York, NY, USA, 2009; pp. 1–5. [Google Scholar] [CrossRef]
  40. Kuwata, Y.; Wolf, M.T.; Zarzhitsky, D.; Huntsberger, T.L. Safe maritime autonomous navigation with COLREGs, using velocity obstacles. IEEE J. Ocean. Eng. 2014, 39, 110–119. [Google Scholar] [CrossRef]
  41. Benjamin, M.R.; Curcio, J.A. COLREGS-based navigation of autonomous marine vehicles. In 2004 IEEE/OES Autonomous Underwater Vehicles (IEEE Cat. No. 04CH37578); IEEE: New York, NY, USA, 2004; pp. 32–39. [Google Scholar] [CrossRef]
  42. Felski, A.; Zwolak, K. The ocean-going autonomous ship—Challenges and threats. J. Mar. Sci. Eng. 2020, 8, 41. [Google Scholar] [CrossRef] [Green Version]
  43. Delvaux, M.; Mayer, G.; Boni, M. Report with Recommendations to the Commission on Civil Law Rules on Robotics, European Parliament. 2017. Available online: https://www.europarl.europa.eu/doceo/document/A-8-2017-0005_EN.pdf (accessed on 1 November 2021).
  44. Leenes, R.; Palmerini, E.; Koops, B.J.; Bertolini, A.; Salvini, P.; Lucivero, F. Regulatory challenges of robotics: Some guidelines for addressing legal and ethical issues. Law Innov. Technol. 2017, 9, 1–44. [Google Scholar] [CrossRef] [Green Version]
  45. Collin, F. Maritime Product Liability at the Dawn of Unmanned Ships—The Finnish Perspective, University of Turku, Faculty of Law Research Paper Series. 2018. Available online: https://www.utu.fi/sites/default/files/public:/media/file/RPR_2_2018.pdf (accessed on 20 January 2022).
  46. Yeomans, G. Autonomous Vehicles Handing Over Control—Opportunities and Risks for Insurance. Lloyd’s 2014. pp. 18–19. Available online: https://www.lloyds.com/~/media/lloyds/reports/emerging%20risk%20reports/autonomous%20vehicles%20final.pdf (accessed on 1 November 2021).
  47. Zipp, J.W. The road will never be the same: A reexamination of tort liability for autonomous vehicles. Transp. Law J. 2016, 43, 137–180. [Google Scholar]
  48. Kim, S. Crashed software: Assessing product liability for software defects in automated vehicles. Duke Law Technol. Rev. 2018, 16, 300–317. [Google Scholar]
  49. Curran, V.G. Cultural immersion, difference and categories in U. S. comparative law. Am. J. Comp. Law 1998, 46, 43–92. [Google Scholar] [CrossRef]
  50. Dannemann, G. Comparative law: Study of similarities or differences? In The Oxford Handbook of Comparative Law; Reimann, M., Zimmerman, R., Eds.; Oxford University Press: Oxford, UK, 2006. [Google Scholar] [CrossRef]
  51. Im, I.; Shin, D.; Jeong, J. Components for smart autonomous ship architecture based on intelligent information technology. Procedia Comput. Sci. 2018, 134, 91–98. [Google Scholar] [CrossRef]
  52. Chatzipanagiotis, M. Third-party liability of manufacturers of automated vehicles for over-the-air updates. In EU Internet Law in the Digital Single Market; Synodinou, T.E., Jougleux, P., Markou, C., Prastitou-Merdi, T., Eds.; Springer: Cham, Switzerland, 2021; pp. 581–603. [Google Scholar] [CrossRef]
Figure 1. Degrees of autonomous mobility between shore and sea from legal and technical perspectives.
Figure 1. Degrees of autonomous mobility between shore and sea from legal and technical perspectives.
Jmse 10 00498 g001
Figure 2. Research questions.
Figure 2. Research questions.
Jmse 10 00498 g002
Figure 3. Research method for legal comparison for MASS.
Figure 3. Research method for legal comparison for MASS.
Jmse 10 00498 g003
Figure 4. Progress of the embedded software error’s cause and risk.
Figure 4. Progress of the embedded software error’s cause and risk.
Jmse 10 00498 g004
Table 1. Overview of liability.
Table 1. Overview of liability.
ItemDefault LiabilityTort LiabilityProduct Liability
Existence of negligenceNecessary conditionNecessary conditionNon-necessary condition
Coverage of damageAll damage or lossAll damage or lossConsequential damage or loss
Table 2. Product liability based on the presumption of defect.
Table 2. Product liability based on the presumption of defect.
ItemProcess of Proving/Presuming Product DefectRemark
Civil Act: Burden of proof of a tortious conduct① Intentional torts/negligence → ② Defect → ③ Infringement of rights or damageComplex
Product Liability Act: Burden of proof of product liability① Defect → ② DamageSimple
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Kim, D.; Lee, C.; Park, S.; Lim, S. Potential Liability Issues of AI-Based Embedded Software in Maritime Autonomous Surface Ships for Maritime Safety in the Korean Maritime Industry. J. Mar. Sci. Eng. 2022, 10, 498. https://doi.org/10.3390/jmse10040498

AMA Style

Kim D, Lee C, Park S, Lim S. Potential Liability Issues of AI-Based Embedded Software in Maritime Autonomous Surface Ships for Maritime Safety in the Korean Maritime Industry. Journal of Marine Science and Engineering. 2022; 10(4):498. https://doi.org/10.3390/jmse10040498

Chicago/Turabian Style

Kim, Daewon, Changhee Lee, Sungho Park, and Sangseop Lim. 2022. "Potential Liability Issues of AI-Based Embedded Software in Maritime Autonomous Surface Ships for Maritime Safety in the Korean Maritime Industry" Journal of Marine Science and Engineering 10, no. 4: 498. https://doi.org/10.3390/jmse10040498

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop