The Bill on Autonomous Shipping: Some Concerns and Suggestions

Keywords: autonomous shipping, liability for harm, vessel’s crew, remote crew.

Autonomous shipping is one of the hot topics of today’s maritime law. Even Russian scholarly literature on maritime law, which does not usually boast a significant number of publications, has provided the reader with a number of works on various problems of autonomous shipping¹.

A bill on autonomous shipping has been drafted and lately introduced to the State Duma of the Russian Federation: the draft federal law “On the Amendment of the Merchant Shipping Code of the Russian Federation and Various Legal Acts of the Russian Federation”² (the “Bill”). Without aiming to present a full-scale analysis of complex, sometimes philosophical, problems of autonomous shipping in general, I would like to concentrate on some features of the Bill and put forward my opinion thereon. This opinion does not pretend to be the ultimate truth and is expressed with the aim of further discussion.

1. Has the time come for the law on autonomous shipping?

The answer to these questions depends on three factors. First, it is the actual state of things in the sphere of autonomous shipping. Second, it is how the problems of autonomous shipping are regulated in private international law. And third, it is the state of domestic legislation and its (un)readiness for the development of autonomous shipping.

I could only comment on the first aspect very briefly and from the position of an average person, as I do not possess the technical knowledge in this field. However, I would probably not be mistaken if I say that the situation is ambiguous. On the one hand, we see develop some significant projects in the field, for example, the launch of the autonomous vessel Yara Birkeland which took place in December 2021. The vessel is to undertake two years of sea trials and is thereafter supposed to start regular operation in autonomous mode. This project, as well as the others of the kind, show that autonomous navigation will be developing. On the other hand, even the technical specialists are now unable to indicate the exact time when the development of artificial intelligence, robotics and adjacent spheres will suffice to implement fully autonomous navigation in the broad sense: not on specially prepared routes or limited areas of navigation, but within the range of the World Ocean or its significant parts. One can only acknowledge that these developments are not yet sufficient. Overall, although AI and robotics play a solid role in our lives, they are still considered as something of an experimental kind and, perhaps, not fully trustworthy.

There are accordingly no doubts that in the future the merchant shipping will be more autonomous than it is now. But when exactly this future comes, we are not aware. In the past, there have been examples of excessive optimism as to the speed of development of various technologies. One could remember the expectations of the fast space exploration that existed in the middle of the 20th century, or the similar hopes of the same age to quickly master machine translation. In reality both processes took a longer time to develop.

Besides, even the most convinced supporters of autonomous shipping acknowledge that the process requires a relatively lengthy transitional period. The majority of vessels currently on the market, and of those being constructed on shipyards, are traditional. Therefore, one can presume that at least for the next 15–20 years a significant part of the world fleet will not be fully autonomous.

As to the second factor, one should pay attention to the work carried out by the International Maritime Organization (IMO). In 2021 it published a scoping paper³ prepared after four-year research aimed at defining the ways to adapt current international maritime law to the needs of autonomous shipping. The way this work is organized leads to an assumption that the IMO does not aim to finalize this immense work in just a few years. The work implies comprehensive, full-scale analysis of applicable rules and, where required, their accurate amendment supported by the results of a thorough discussion. Apparently, the quality of this work is more important than its speed.

In this situation, Russian legislator appears to have the following choice. The first option is to wait until the IMO finalizes its work and to reform domestic legislation in compliance with the results of this global project. The second option is to elaborate on its own solutions without waiting for IMO’s conclusions. One may assume that currently the second approach prevails. Its advantage is the opportunity not to wait until the massive and, apparently, lengthy project of the IMO comes to its conclusion. But, in my opinion, it also has significant drawbacks. The domestic legislator can produce a piece of national regulation on autonomous shipping. But application of such regulation outside Russian jurisdiction will be impossible. Besides, at some point of time the project of the IMO will be brought to its conclusion, which will imply amendments to international conventions and IMO documents that are currently based on the necessity to have crew on board⁴. At this point of time Russian legislator will in any case have to amend domestic rules so that they comply with the (newly amended) international regulation. Is it worth developing a domestic regime of autonomous shipping now, having in mind the necessity to amend it in future once again?

An intermediate option is also available: to start the work on the domestic legislation right away, but to take into consideration the recommendations that have been elaborated, albeit on a preliminary basis, by the international organizations and primarily by the IMO. It will be shown below that the Bill does not always follow this path.

Turning to the third factor: on 5 December 2020 the Government has issued a Ruling No. 2031 “On the Experimental Operation of Autonomous Vessels under the Flag of the Russian Federation” (the “Ruling”). The experiment is taking place from 2022 till 2025. It allows shipowners who are interested to operate autonomous vessels to do it legally, although subject to certain conditions. In particular, a vessel that participates in the experiment may obtain a “reduced” minimal crew certificate that would take into account the level of the vessel’s autonomy.

At the same time the Ruling does not in fact regulate the issues of private law that arise in connection with the operation of autonomous vessels. It limits itself to providing for shipowners’ obligations to ensure the safety of the operation of autonomous vessels and to provide their counterparts (consignees, insurers etc.) information on the participation of the vessel in the experiment. Therefore, the issues of contractual or tortious liabilities that may arise in connection with the operation of autonomous vessels are regulated by general rules of the Merchant Shipping Code and civil law. These rules are based on a presumption that the vessel is equipped with a crew and, in particular, has a master. They charge those persons with particular obligations and provide liability for the failure to comply therewith: see, for example, paragraph 1 of article 144, paragraph 3 of article 153, article 198, article 202 of the Merchant Shipping Code. The application of these rules to autonomous vessels requires some adaptation. On the other hand, a significant advantage of the legislation in force is that it sets out clear rules and approaches to the issues of contractual and tortious liability.

To summarize, there are serious doubts that the law on autonomous shipping has to be enacted right away. In my opinion, it would be wiser to do it later, when some experience of autonomous shipping is gathered. Even if we proceed on a presumption that the legislation on autonomous shipping should be elaborated without delay, the speed should not impair quality which, in turn, implies the consideration of international law and experience and the need to find appropriate solutions to the issues of autonomous vessels’ liability. Unfortunately, the Bill is not always up to these tasks.

2. The analysis of the Bill

As already mentioned, this analysis does not purport to be comprehensive. I would like to point at only some of the peculiarities of the Bill that, in my opinion, should be subject of a further discussion.

2.1. Terms used by the Bill

The IMO Scoping Paper is based on the classification of autonomous ships into four degrees of autonomy. Degree one encompasses ships with seafarers on board where processes and decision support are automated. Degree two – remotely controlled ships with seafarers on board. Degree three – remotely controlled ships without seafarers on board. And degree four: fully autonomous ships.

The Bill does not use this classification, introducing instead two categories of autonomous vessels: semi-autonomous and fully autonomous ships. In my opinion, the classification used by the Bill raises a number of serious concerns.

The definition of a semi-autonomous vessel that is provided by the Bill does not allow to distinguish it from adjacent types of vessels. As per the Bill, a semi-autonomous vessel “can sail without permanent control… and without permanent handling of navigation”. Her crew “handles general observation” and interferes into the vessel’s operation “in case of necessity”.

I believe that the term “without permanent control” is unclear. Almost all modern vessels have some of their functions automated. The crew members control that the automatic machinery functions properly and interfere into their work in case of necessity. So, where is the distinction between an average modern vessel and the “semi-autonomous vessel” as per the Bill?

The definition of fully autonomous vessels is also inaccurate. It reads as follows: “a fully autonomous vessel… can sail… without permanent observation [by a remote crew] and without control of navigation”. It is unclear from this definition whether a fully autonomous vessel may operate without crew at all, or she still requires some control on behalf of the crew, although in the remote mode. If we suppose that the first way of construction is correct, i.e., that a fully autonomous vessel may operate without human intervention at all, then it is very doubtful to join two very different types of vessels under the same title “fully autonomous vessel”: a vessel that is operated remotely and a vessel that operates without human intervention. These situations are apparently different and require different regulation of liability issues.

The above concerns may seem to have a technical nature. However, I believe there are no trifles in such issues. We see a new legal regime for some types of vessels, defining such important issues as requirements to safety of navigation, tortious liability of the vessel, labour relations and the status of remote crew etc. Therefore, it is important to define clearly which vessels fall under this regime and which do not.

2.2. The crew of a fully autonomous ship

One of the important questions that arise in any project related to autonomous shipping is the status of personnel who operate the vessel not being on board. The initial version of the Bill provided that the vessel in this case has no crew. During the subsequent work on the Bill its developers took into account the proposal made by the Council for Codification of Civil Law⁵, and the Bill currently refers to such personnel as to “remote crew”. I believe this is a correct amendment. It allows to apply to the employers who in fact operate the ship, although not being on board, such higher standards that are required of a vessel’s crew. Such standards are provided in articles 106.5 and 106.8 that the Bill purports to introduce into the Merchant Shipping Code.

This decision also has a drawback: if the remote personnel are considered to be crew members, they may claim the benefits arising out of specific guarantees applicable to seafarers (such as specific labour conditions etc.)⁶. It will be up to the courts to consider if, and to what extent the remote crew, who do not appear to be subject to maritime perils, may claim the application of specific seafarers’ guarantees. In my opinion, the work of the remote crew is not connected with dangers and inconvenience that is typical for traditional seafarers’ duties at sea and therefore should not be protected by seafarers’ labour guarantees. However, because of the vessels are potentially dangerous for third parties as well as the environment, the same qualification requirements that apply to seafarers should apply to the remote crew members⁷.

The position of the IMO is that determining the legal status of the captain, crew members, persons in charge of the operation of the autonomous vessels is a key task of the project. The Scoping Paper proposes to consider two opposite approaches: to acknowledge that remote crew members are seafarers or to refuse to do it⁸. At this moment both approaches are being scrutinized, and it is yet impossible to say which of them would prevail.

2.3. Operation of the autonomous vessel

The most important provisions related to the navigation of autonomous vessels can be found in chapter VI.1 which is proposed to be introduced into the Merchant Shipping Code. There is a number of discrepancies in this chapter that might seem to be technical, but still may lead to serious problems. For example, paragraph 1 of article 106.1 says that the operation of the ship (including navigation, ensuring that safety measures are being kept up with etc.) is carried out under the supervision of the vessel’s crew (if it is available) and under the supervision of the remote crew. The use of “and” means that a semi-autonomous vessel must be under simultaneous supervision of the on board crew and of the remote crew. But other provisions of the Bill provide that the vessel controlled by the remote crew is a fully autonomous vessel and therefore does not have an on board crew⁹.

Another discrepancy of this kind may be found in paragraphs 2 and 3 of the newly drafted article 106.1 of the Merchant Shipping Code. Paragraph 2 says that the shipowner must ensure continuous supervision of the autonomous vessel. Paragraph 3 contains an exclusion from this rule: a fully autonomous vessel that is fit for independent navigation, is relieved from requirement of continuous supervision and operation. But what is the distinction between the ships that fall under paragraph 2 and paragraph 3, respectively? Who defines whether the level of autonomy is sufficient to obtain relief from the obligation of continuous supervision? The same article of the Bill mentions the “Rules of the Operation of Autonomous Vessels” (the “Rules of Operation”) that are to be adopted by the Ministry of transport. But as for now these rules do not exist. Besides, strictly speaking, paragraph 3 of article 106.1 does not say that only such vessels whose autonomy level is up to the criteria that are to be set in the Rules of Operation are relieved from the supervision by the remote crew. Therefore, the question as to which vessels must be supervised and operated by the remote crew and which are not under this obligation, is still open.

2.4. Liability for harm caused by the autonomous vessel

In my opinion, one of the most significant drawbacks of the Bill is that it practically does not regulate liability. It is easy to notice that both foreign and domestic research on autonomous shipping is focused on the issues of liability¹⁰. Indeed, the liability that arises in connection with the operation of autonomous vessels may differ from the liability for damages caused by traditional vessels. Firstly, as I will show below, the Bill contains inaccurate and contradictory provisions on the liability for harm caused by the vessel. Here, in my opinion, we see a drawback of Bill and not the inherent problem of the autonomous shipping. Secondly, according to the Bill the remote crew of an autonomous vessel can be hired and controlled by a so-called “competent organization”. Its status differs from the roles that are well known in maritime law. At least, it does not follow from the Bill that the “competent organization” carries out the function of the ship’s operator, or any other renown function. Therefore, the remote crew is controlled by someone whose status is unclear. Having said this, we should recall that under the general rule set out in article 1068 of the Russian Civil Code, where an employee (i.e., remote crew member) is guilty of damage, it should be compensated by the employer (“competent organization”). However, the Bill does not contain any provisions as to the liability of the competent organization¹¹.

Thirdly, as foreign researchers point out, the liability in connection with the operation of an autonomous vessel may arise not only with the shipowners but also with the developers of the vessel and/or the software that ensures her independent navigation. Usually, the liability of such persons is discussed on the basis of the general rules of product liability (articles 1096-1098 of the Russian Civil Code). However, the Bill does not say anything about this issue, either.

Generally, the provisions on liability contained in the Bill are scarce. One could name paragraph 4 of article 106.1, articles 106.7 and 106.8. Under paragraph 4 of article 106.1 “the shipowners of an autonomous vessel shall ensure safe operation of the vessel and her seaworthiness”. This paragraph raises a number of questions. For example, does the fact that only the shipowners are mentioned as subjects of liability mean that the liability of any other persons is excluded? Article 106.7 provides that the liability is carried by “shipowners and [registered] owners”. This wording does not clear the doubts as to the potential liability of other parties. Besides, shipowners and registered owners could be different persons. These discrepancies in the Bill’s wording do not allow to clearly understand who is going to be liable – the registered owners where the international treaty and/or the Merchant Shipping Code nominates them as the party liable for damages, and the shipowners in all other cases, or is the liability of the shipowners and the registered owners joint and several.

Taking into account the serious risks associated with the operation of autonomous vessels, the question of liability should be in any case resolved unequivocally. Persons against whom claims should be brought by injured parties must be clearly defined; where there is more than one defendant, it should be made clear whether their liability is joint or subsidiary.

Paragraph 2 of article 106.8 of the Merchant Shipping Code (as proposed by the Bill) is, in my opinion, fully wrong. It reads as follows: “the member of the remote crew… while navigating the vessel… is the person in charge of the navigation… and shall perform all functions and requirements with which the master is charged according to the legislation and international agreements”. First of all, this provision contradicts the principle of sole liability of the captain for the navigation of the vessel. The Bill amends article 61 of the Merchant Shipping Code where this principle is stated, providing for an exception. Nevertheless, it should not result in uncertainty as to which remote crew members are in charge instead of the captain. Secondly, the provision can be understood as stating that the remote crew member is liable for the damages done by the vessel, i.e., that the crew member should be a defendant upon a claim brought by the injured parties. In fact, other parties are clearly liable to third parties: those are the registered owners and/or the shipowners; direct claims may be brought against the liability insurers; claims against developers of software and against the “competent organizations” may also be discussed. After these primary defendants compensate the injured parties, they may have a right of redress towards the remote crew members¹².

2.5. Carriage of goods by autonomous vessels: some peculiarities

Apart from liability, the Bill is also silent on issues related to the carriage of goods by sea¹³. For example, how will the cargo be loaded without participation of the crew? In what manner the carriers who do not have a crew on board will perform their obligation to “properly and carefully handle, stow, keep, care for, and discharge the goods carried” (article 150 of the Merchant Shipping Code)? How would the bill of lading be signed and issued? One may suppose that, where there is no crew on the vessel, this task will be performed by the remote crew, and the bill of lading will be issued electronically. But it is evident that there will be some peculiar issues that do not arise where cargoes are loaded under the supervision of people.

Similar questions may be asked in relation to potential extraordinary situations with the cargo during shipping. What if the cargo is on fire, or shifts so that it threatens the vessel’s stability? The legislation currently in force provides that in such situations the crew must perform all possible measures to eliminate the danger. The actions of the crew are then analyzed to determine whether the carriers have duly performed their obligations. How would these problems be resolved where there is no crew on board? The Bill remains silent on this issue.

2.6. Liability insurance of autonomous vessels

As the Bill was revised, its authors inserted thereto a provision on liability insurance (paragraph 2 of article 106.7). On the one hand, this provision is an important amendment. On the other hand, it is clearly insufficient and does not resolve the problem, as it contains no details and only refers to the unnamed act of Government that is yet to be drafted.

This is inappropriate, as liability insurance is even more required for autonomous vessels than for “traditional” ones. Where damage is done by a traditional vessel, the injured party may protect itself by arresting the vessel. It is unclear how an arrest will look like where the ship is operated by a remote crew or navigates without a crew at all, being programmed for a particular route. Evidently, detention and arrest of ships will be performed in a different manner. Besides, when a traditional ship is arrested, one of the factors that stimulates the shipowners to deal with the problem is the need to support the crew, or to lift it off the vessel and repatriate. This factor is no longer effective where a ship without a crew is detained.

The shipowners and the remote crew may be located very far from the place where the autonomous ship is arrested. There are currently no international conventions on autonomous shipping, and the Bill may apply only within Russian jurisdiction. But it directly allows foreign-flagged autonomous vessels to call Russian ports (article 16.2 of the Merchant Shipping Code as amended by the Bill). If a foreign vessel causes damage in Russia, Russian claimants may have difficulties in reaching foreign defendants. Besides, both the foreign and the Russian shipowners may go bankrupt. In these cases, plaintiffs whose interests were adversely affected by an autonomous vessel will most likely be unable to receive compensation.

Surely, the vessel herself acts as a security for potential claims. But its costs may be insufficient to make good all claims. Besides, the procedure of forced sale and the handover of an autonomous vessel to a new owner is quite unclear. Where the vessel is operated remotely, the physical handover of the hull will not be sufficient to transfer full control over the vessel.

Because detention, arrest and forced sale in relation to autonomous vessel are expected to face additional difficulties, liability insurance becomes even more important. Therefore, the provision of paragraph 2 of article 106.7 of the Bill, devoid of any particular details or requirements to such insurance, is clearly insufficient.

2.7. Some additional comments to the Bill

There is also a number of comments to the Bill, which, although have a considerable meaning, belong to various issues and problems.

The Ruling that currently regulates the operation of autonomous vessel in Russia provides for informational duties of the shipowner of autonomous vessel. Indeed, considering the experimental nature of autonomous shipping the owners’ counterparts (shippers, tug owners etc.) should be able to receive additional information. It should include at least information on the level of autonomy, on the method of communication with the owners, on the procedures in emergency situations, on the peculiarities of shipping documents. This issue is, however, not covered by the Bill.

Article 106.2 of the Merchant Shipping Code (as amended by the Bill) provides that “in case of military activities… or other cases of military danger the shipowners shall take all measures to prevent loss, damage or seizure of the vessel”. The Bill does not further identify which measures may be taken under such conditions by the remote crew which may be located on a significant distance from the vessel. Having said that, I believe that many emergency situations that still occur on marine transport require personal decisions, above all the decisions of the master.

Under article 106.3, shipowners and cargo owners represent their interests independently. On may assume that by this wording the authors of the Bill decided to bypass the provisions of the Merchant Shipping code on the functions and authorities of the captain, in particular, the right to sign a contract of salvage on behalf of shipowners as well as cargo owners. Where cargo is carried on the autonomous vessel, how would the procedure of entering into a salvage contract look like under these new provisions? Will it be feasible at all, taking into consideration that a salvage contract should be signed promptly, while the shipowners and cargo owners may be located in various jurisdictions and have various opinions on the salvage conditions? The salvage operation encompasses not only the interests of shipowners and cargo owners, but also of third parties to whom environmental damage may be done. But this issue is not regulated by the Bill, either.

The expert opinions of the Council of Codification pointed out that the Bill does not regulate the issue of licensing and use of software that enables operation of autonomous vessel. The Bill does not set any requirements to such software, either. The current wording of the Bill continues to remain silent on the issue. It is however clear that the use of software, requirements to software and the issues of liability of software developers are key problems that differ autonomous vessels from the traditional ones. They have their own characteristic features and will hardly be resolved without statutory regulation.

Conclusion

The wording of the Bill considered in this article is still subject to change. The comparison between the previous wordings and the current one shows that the authors of the Bill already carried out a significant work that made the provisions of the Bill more accurate. Still, a very important part of the work is still ahead. When doing this work, one should above all take into consideration the interests of the marine industry. In my opinion, it needs clear and effective guarantees of safety of navigation. Legal and financial mechanisms of compensation of potential damages that may be done during operation of autonomous vessels should be secured, too. Without a comprehensive answer to the above concerns the Bill may bring more harm than good. As international experience shows, when finding the solutions for complicated issues such as shown above it is more important to undertake a comprehensive research and find legally accurate decisions than to proceed at full speed. ■

Notes