2.2.1 Definition of Autonomous Ships and Analysis of Levels of Automation in Ship Operation

The notion of autonomous vessels remains notably elusive, with no singular, universally recognized definition to command the field of discourse. However, the definition posited by the IMO is one of the most widely acknowledged and frequently cited. The IMO opts for the term “Maritime Autonomous Surface Ship” (MASS), meaning “a ship which, to a varying degree, can operate independently of human interaction” (IMO 2021, annex, 3, 2018a). The IMO further refines this definition by categorizing MASS into four distinct classes, gradated according to the extent to which automation permeates the operational functions of the vessel.Footnote ²⁰ In this categorization,Footnote ²¹ with ships in the first degree of autonomy, the crew is on board for the purpose of providing the ship’s direction and management. However, there is automation in some operations and the seafarer can intervene when necessary. A ship in this group is referred to as a “ship with automated process and decision support.” Ships with a second degree of autonomy can be controlled and operated remotely. However, the crew continues to be on board to make interventions when necessary. The ships that will bring about the most radical change in the maritime industry are those with the third and fourth degrees of autonomy, due to the absence of a crew on board the ship. In third-degree ships, the crew is not on board; it is accepted that the ship is operated and managed remotely by personnel in a center established on land or on shore. A ship in this category is a “remotely controlled ship without seafarers on board,” or a semi-autonomous ship. Finally, ships with a fourth degree of autonomy, or “fully autonomous ships,” have no crew on board and no active control by personnel from the remote operation center. Through AI systems integrated into the system, the ship determines its movement by evaluating the current situation and conditions and making the necessary decisions regarding the operation of the ship without the need for human support. In spite of the self-operation feature of this group of ships, remote operation centers still exist,Footnote ²² but the role of center personnel is limited to monitoring the actions of the system and decision-making process in case of system failure (IMO 2017a, 6). Whatever the method of operation and the level of autonomy, there should be a person in charge of the autonomous ship, either on board or in a remote operation center, who has the authority to intervene, performing the duties of the master (IMO 2023e, 42, 2023b, 4). Regarding the relationship between these accepted degrees, the IMO suggests that the levels do not represent a hierarchical order (IMO 2018b, annex 1, 1, 2021, annex, 4).

Based on these autonomy levels, theoretically, in a particular case, a ship may operate at just one level. There may be no expectation that the ship’s autonomy level will change; the ship is assumed to operate at a fixed autonomy level throughout the voyage, or to maintain a “static autonomy level” (Barber et al. Reference Barber, Goel and Martin2000, 145). However, it is also possible to design systems where the autonomy level changes during the voyage, based on different conditions or operations. For example, a ship may be designed to operate autonomously while sailing on the high seas, but the design might also enable control to transfer to the operator or crew during specific phases of the voyage, such as passing through a busy strait, entering a port, or performing docking or loading operations (AAWA 2016, 30; Ringbom and Collin Reference Ringbom, Collin, Ringbom, Røsæg and Solvang2021, 14–15). A ship’s autonomy level may change multiple times during the voyage (Barber et al. Reference Barber, Goel and Martin2000, 130; Sözer Reference Sözer2024, 45). This is referred to as a “dynamic autonomy level” in ship operation. Autonomous ships will transition freely between different levels of autonomy during a particular voyage (AAWA 2016, 6–7; Fukuto Reference Fukuto2021, 47; Maritime UK 2022, 20–21; Ringbom and Collin Reference Ringbom, Collin, Ringbom, Røsæg and Solvang2021, 14–15; Yang et al. Reference Yang, Utne, Sandøy, Ramos and Rokseth2020). The operation and design of autonomous ships currently in service, such as the Yara BirkelandFootnote ²³ and the Prism Courage,Footnote ²⁴ support this submission.Footnote ²⁵ The analysis in this chapter is approached through the prism of “dynamic autonomy.” The term “autonomous ship” is used to refer to a ship operated without any crew on board.

2.2.2 Remote Operation Center

Unlike conventional ships, autonomous ships consist of two separate parts, the first of which has a classical appearance at sea. Because the crew is completely separated from the ship, its operation and monitoring involve a remote operation center.Footnote ²⁶ That center is defined as “a location remote from the MASS that can operate some or all aspects of the functions of the MASS” (IMO 2023b, 6).Footnote ²⁷ The center is also defined as a structure consisting of computer hardware, telecommunications equipment, and various high-tech tools arranged to perform a specific function, managed and operated by the relevant personnel (Sözer Reference Sözer2024, 84).Footnote ²⁸ It is possible to list the functions of the center as monitoring the ship, ensuring the ship’s direction and management, collecting data, and fulfilling certain legal obligations (Sözer Reference Sözer2024, 89–92). However, depending on the fact of each case, its main function may change during a particular voyage. Following the dynamic autonomy submission, for example, if the ship proceeds on the high seas, the main function of the center might be to monitor the ship. But, if the ship is sailing through a busy strait, the main function of the center might be to control the ship. Physical conditions and technical conditions may require such an active role change by the center during the voyage. It may not be technically possible to sail the ship remotely throughout the voyage – for example, where continuous data exchange and communication between the ship and the operation center, located either on land or aboard another vessel,Footnote ²⁹ is not provided (AAWA 2016, 4). This communication would depend on the transmission of high-resolution and large-scale data collected by various sensors, cameras, radar, infrared technology, and the Automatic Identification System (AIS), ensuring that the operation center can maintain uninterrupted access to the ship’s systems (Ciğer Reference Ciğer2024, 1145).Footnote ³⁰ The reliability of current technologies to facilitate such data transfer is debatable, particularly concerning potential risks arising from connection delays, as well as adverse factors such as weather and distance (1145).Footnote ³¹ Solutions may involve utilizing both technologies (remote operation and autonomous operation) in tandem (1146)Footnote ³² and active role changes at the remote operation center. The ship will be controlled through an interface in which the bridge is simulated in the center, which replicates camera and sensor systems on the ship, plus controls such as steering gear and joysticks for the remote operator.Footnote ³³

There are three main issues around the concept of the remote operation center. The first relates to the status of the remote operator.Footnote ³⁴ The remote operator is responsible for monitoring the ship’s condition, course, speed, and environmental factors during navigation in light of the information sent from the ship; supervising the decision held by the system; and, when necessary, taking direct operation and management of the ship (Ciğer Reference Ciğer2024, 1149; Sözer Reference Sözer2024, 88–93). The Joint Working Group (JWG) has yet to explain fully the role of the remote operator, though it has provisionally affirmed that, irrespective of the level of autonomy, a human captain must preside over each autonomous ship (IMO 2023b, 4). In the report submitted by the International Federation of Shipmaster’s Associations on the roles and responsibilities of the master, it is also emphasized that autonomous ships are ships by nature and that they need a master. Autonomous ships, like all ships, should be under the command and control of a master who has both situational awareness and sufficient maritime knowledge (IMO 2023a, 4). The role of the ship master will not disappear, but the execution of the role may be filled by a person on land who has the necessary qualifications and can exercise discretionary authority when necessary (4). A MASS ship master may be preferred over traditional manning (IMO 2023a, annex, 1). At this point, it should be mentioned that the UK was among the first countries to issue regulations on personnel working in remote operation centers. The Maritime and Coastguard Agency adopted the Workboat Code Edition 3 in December 2023. The technical requirements for the operation of ships remotely controlled are regulated within Annex 2. The Code is also important in terms of the definition of the remote operator. In Annex 2, the remote operator is defined as “any person, including the Master, with recognised or certifiable experience who is engaged in the remote operation of a [Remotely Operated Unmanned Vessel]” (MCA 2023, 27). The existence of a master in the remote operation center is also accepted by the JWG, which notes that the human master should have the means to intervene in ship operations when necessary, regardless of the mode of operation and the autonomy level (IMO 2023b, 4). Analysis held in the literature also demonstrates that the general trend is to treat the remote operator of the ship as a master while the application of master-related terms is evaluated. For example, the authors who analyze the application of Article 4(2)(a) of the Hague-Visby RulesFootnote ³⁵ to autonomous ships submitted that a remote operator can be treated as the functional equivalent of a master in the context of the relevant article, so that the carrier can escape from liability for loss and damage arising due to the fault of the remote operator in the navigation or management of the autonomous ship (Baughen Reference Baughen, Soyer and Tettenborn2021, 87–88; Stevens Reference Stevens, Soyer and Tettenborn2020, 159, 161). The prevailing view, in essence, is to regard those entrusted with the control and operation of the ship – despite their absence from the ship’s physical presence – as holding positions akin to master and crew.Footnote ³⁶ It follows that personnel tasked with the governance of autonomous ships should be recognized in a capacity resembling that of the ship’s master, bearing responsibilities comparable to those traditionally assigned to the master. This rationale, therefore, warrants the assertion that the relationship between the shore-based operational center staff and their autonomous vessels mirrors, in substance, the long-established association between the master and the time charterer.

The second issue regarding the remote operation center is whether each center and remote operator will be responsible for the operation of only one ship or more than one. The MUNIN project plans for up to ninety ships controlled from a remote center (Porathe et al. Reference Porathe, Costa and Tjora2014, 2). The project also suggests that one operator may be responsible for monitoring and controlling up to six ships (Man et al. Reference Man, Lundh and Porathe2015, 2676). The possibility of a master being in charge of more than one ship was supported by the JWG (IMO 2023b, 5). It can be inferred from this approach that the general view of the JWG is that more than one ship can be controlled from one center. One of the currently active remote operation centers, Massterly in Horten, Norway, provides management and operation services for the Yara Birkeland and ASKO’s electric barges MS Marit and MS Therese. This example also demonstrates the practical application of a single center controlling multiple ships (Massterly 2024; Wilhelmsen 2024). In fact, if it is accepted that each vessel will be controlled from a separate center, this will require a huge number of centers to be set up and staffed. From an economic perspective, this proposition appears to lack rationality, so probably the most effective approach would be to operate as many ships as possible from a single center.Footnote ³⁷

The final issue is who will provide the remote operation service. It can be argued that the owner of an autonomous ship can construct its own center and staff it with personnel to monitor and operate the ship. However, these centers consist of computer hardware, telecommunications equipment, and various high-tech tools. The cost of establishing a remote operation center varies based on factors such as facility size, technology integration, and operational requirements. For example, while the cost of equipping and maintaining an autonomous bulker, inclusive of annual operational expenses and rental commitments, remains a matter of rough estimation, it is acknowledged that the investment for the development of the remote operation center lies between US$1 million and US$2.1 million, with annual operational outlay approximating US$0.87 million. The establishment of such a center entails a recurring cost of US$33,000 per vessel annually (Ziajka-Poznańska and Montewka Reference Ziajka-Poznańska and Montewka2021, 6–8), alongside the considerable infrastructure expenses, as exemplified by the KRISO research center’s construction cost of 22.1 billion Korean won (approximately US$16 million) (KRISO n.d.). Following these values, the construction of an individual center by each owner may not be considered economically viable, especially if the shipowner operates a single ship. Since, as noted above, operating more than one ship from the center is a sensible choice, it could be argued that where the shipowner operates larger fleets, it could set up its own remote operation center and all its ships could be operated from the same center. Considering the fact that shipping companies even arrange special ports for their fleet,Footnote ³⁸ this could also be done with respect to the operation center and it could be cost-effective. However, depending on the area it is traveling through, as the ship moves from one port to another it may be necessary for it to change the operation center (Sözer Reference Sözer2024, 85; see also Baughen Reference Baughen, Soyer and Tettenborn2021, 84),Footnote ³⁹ such as the air traffic control area and the vessel traffic services system.Footnote ⁴⁰ This may be due to the difficulty of transmitting information when the ship is at a great distance from the center (Sözer Reference Sözer2024, 85; see also Carey Reference Carey2019, 7).Footnote ⁴¹ When the ship goes outside the control area of the remote operation center, the owner of the ship might be required to connect with another center. Rather than the service of remote operation being provided by the shipowner and associated maritime centers throughout the entire voyage, third-party shipping management companies will usually be involved in the operation process. It is not uncommon for shipowners that do not have shipping knowledge and experience to contract with a company to provide technical, crew, and commercial management and ancillary services (Plomaritou and Papadopoulos Reference Plomaritou and Papadopoulos2017, 115). The company can provide a remote operations center and personnel to operate the ship, either onshore or on board, as well as serve as the technical managers of the ship (Hunter Reference Hunter2020).Footnote ⁴² Steps taken by the Baltic and International Maritime Council (BIMCO) actually reflect this view. The AUTOSHIPMAN agreement was recently adopted by BIMCO in view of the increasing number of remotely operated ships, with the aim of providing a contractual framework for third-party ship managers that deliver operational services for these ships. It contains terms regarding the obligations, responsibilities, and liabilities of the parties (BIMCO 2024b). The form is drafted considering the SHIPMAN 2024 agreement (BIMCO 2024a). In addition, the number of companies offering management and operation services for remotely operated and autonomous vessels is growing by the day (Massterly n.d.; Ocean Infinity 2022). The specialized nature of the capabilities and technology featured in these remote operation centers means that it is prudent to delegate ship responsibilities to specialized entities.

2.4.1. The Shipowner’s Obligation to Provide a Ship That Complies with the Charter Description at the Time of Delivery

It is common to see a clause in time charters that describes features of the ship such as class, cargo capacity, nationality, speed, and bunker consumption. From the charterer’s point of view, this clause is probably one of the most important parts of the time charter, as the charterer agrees to pay a fixed rent for an unknown vessel for the entire period of the charter, based solely on the contractual description of the vessel. The shipowner’s obligation under the description clause is to provide a ship that conforms to the contractual description (Coghlin et al. Reference Coghlin, Baker, Kenny, Kimball and Belknap2014, para. I.33). Providing a ship that does not comply with the charter description is a breach of the charter, and the charterer is entitled to damages for the loss suffered, unless the charterparty contains a provision that excludes the shipowner’s liability for its failure (para. 3.18). If the description in the charter indicates condition status,Footnote ⁵⁴ such as the ship’s class,Footnote ⁵⁵ or an intermediate term status, such as the ship’s speed and bunker consumption (Coghlin et al. Reference Coghlin, Baker, Kenny, Kimball and Belknap2014, para. 3.77),Footnote ⁵⁶ the shipowner’s failure to provide a ship that complies with the description can be considered a repudiatory breach, and the charterer has a right to terminate the charter. Under NYPE 2015, descriptions as to the ship are given in the preamble. In addition, the form contains Annex A (“Vessel Description”) – a detailed vessel questionnaire that forms part of the charterparty. There is also Clause 12, under which speed and bunker consumption are stipulated.Footnote ⁵⁷ When time charters are made for autonomous ships, the descriptive information given about the ship in the form needs to be extended.

Autonomous shipping possesses distinctive features that will facilitate functions such as mooring, berthing, and cargo handling without the need for human assistance. The Yara Birkeland, a ship that is currently expected to perform these operations, serves as a tangible example of this emerging technological development.Footnote ⁵⁸ From the charterer’s perspective, capabilities of the ship to perform these operations, especially cargo handling, might be important. Under Clause 8 of NYPE 2015, “charterers shall perform all cargo handling, including but not limited to loading, stowing, trimming, lashing, securing, dunnaging, unlashing, discharging, and tallying, at their risk and expense, under the supervision of the Master.” If the ship is capable of performing cargo-handling autonomously, this provision will be inapplicable. Since the charterer is not the one who has the right of ownership or possession of the chartered ship, it would not be reasonable to hold it liable for cargo-handling, which is performed by the ship under remote or autonomous operation without involvement of the charterer. In such a case, it will be more proper to impose this obligation on the shoulders of the shipowner. Thus, an autonomous ship’s cargo-handling capability requires a profound shift in terms of the liability. Therefore, if the autonomous ship has such a capability, the charterer will want to be aware of this information, so the description clause should be extended by making this information a part of the ship’s description.

It might also be important for the charterer to know the location of the remote operation center. Although the remote operation center and the ship are separate elements, without the center the ship will not have any meaning for the charterer. Therefore, the time charter is now a contract for the use of the ship with an associated remote operation center and personnel. This author suggests that information related to the capacity of potential remote operation centers arranged for the ship can be made a part of the ship’s description. For example, if the shipowner prefers an operation center that has limited capacity and only allows the ship to be used within the geographical limit of a particular region due to data transmission difficulties and if the shipowner does not have any intention to arrange another center that extends the employment area of the ship outside of the specified region, this information should be given as part of a description clause and/or under Clause 1(b) as part of the trading limits.Footnote ⁵⁹ It is important that the charterer knows this beforehand, because the information is relevant to plan future employment. If its intention is to use the ship outside of the designated area, it might not enter into the charter agreement for that ship. There might also be some cases in which the shipowner changes the remote operation center during the period of charter. The reason behind this may be due to problems the shipowner encounters with the remote operation center or its closure. The shipowner should be careful about such a change. One of the implied obligations imposed on the shipowner under the time charter is that the shipowner will not, at any stage of the charter, make any alterations to the ship that may (or potentially will) affect the services to be provided.Footnote ⁶⁰ Following this, if the remote operation center is changed during the charter period, the shipowner must arrange a center that has at least the same capabilities as the previous one. If we consider this in terms of the geographical area in which the operation center can provide the service, the center that limits the charterer’s area of operation compared to the previous center should not be chosen by the shipowner. Otherwise, this will be a breach of the shipowner’s implied obligation, and the charterer is entitled to damages. If it is proved that this alteration goes to the root of the contract, the charterer may also be entitled to terminate the charter.Footnote ⁶¹ Some information as to personnel in the remote operation center can also be important for the charterer to evaluate the seaworthiness of the autonomous ship. However, this information is more accurately considered under the clause dealing with the seaworthiness of the ship rather than under the description clause.

Under NYPE 2015, some changes might also be needed related to the description given to warranties for speed and bunker consumption (performance). These warranties are usually subject to weather conditions.Footnote ⁶² This means that the shipowner warrants that the ship will be able to achieve the described speed and bunker consumption only if good weather conditions exist. Good weather is commonly defined in time charters referring to the BeaufortFootnote ⁶³ and Douglas scales.Footnote ⁶⁴ For example, Clause 12(a) of NYPE 2015 refers to both scales to describe good weather as “wind up to and including Force four (4) as per the Beaufort Scale and sea state up to and including Sea State three (3) as per the Douglas Sea Scale.”Footnote ⁶⁵ In a particular case where the time charterer asserts that the ship underperformed during the voyage, the burden of proof rests on the charterer. In order to clear this hurdle, the charterer needs to show that, although good weather conditions existed, the ship underperformed. To prove this, the chartering party normally relies on a report of the ship’s performance prepared by a weather routing company (WRC). Clause 12(b) of NYPE 2015, which states that “the Charterers shall have the option of using their preferred weather routing service,” provides an option to the charterer. It is made clear later, in Clause 12(d) of NYPE 2015, that the charterer can rely on this report to support its claim for underperformance. On the other hand, the ship’s logbook prepared by the master is preferred by the shipowner as a main source of evidence to rebut the chartering party’s underperformance claims in a particular case.Footnote ⁶⁶ Since they are prepared by different parties, there are often inconsistencies between weather and sea conditions recorded in a WRC report and the ship’s logbook. Generally, the records maintained by the master in a logbook are given greater credence than those recorded in a WRC report. This perspective appears to be logical, given the master’s status as a trained weather observer recognized by the World Meteorological Organization. The master’s presence on board the vessel during the entire charter period, observing weather and sea conditions firsthand, further substantiates the credibility of these records (Harris Reference Harris2014, 4). This conventional view also supports the ship’s logbook by pointing out that the WRC report is prepared by using satellite imagery, which is just a snapshot of weather and sea conditions, and so that may not capture the local impact of weather and sea conditions, such as currents, on the ship’s performance (4). Furthermore, data for the preparation of the report is collected only once or twice a day (Alexander Reference Alexander2013). Where autonomous ships are used, the traditional view will lose its dominance because information in a ship’s logbook entriesFootnote ⁶⁷ on the effect of sea and weather conditions on the ship’s performance will no longer be recorded by the master who is on board the ship but by a person in the remote operation center, or perhaps by AI using shipboard sensors. When autonomous ships are operated under time charter, there will be no difference between the ship’s remote operation center logbookFootnote ⁶⁸ and the WRC report in terms of the way data is collected. Both will be done remotely, but the first will be more reliable as it will be collected by an advanced system of autonomous ships and during the actual voyage. There may even be no need for the charterer’s WRC arrangement. Autonomous ships will reduce the charterer’s main concern that the master may exaggerate the sea and wind conditions and make inaccurate logbook entries to the shipowner’s advantage. Even if the records are kept by personnel working at the center arranged by the shipowner, and there may be some contractual relationship between them and the shipowner or between them and a third-party manager arranged by the shipowner, the center personnel or its AI systems will make entries in the ship’s logbook based on data collected continuously from the ship. Therefore, even if the charterer has concerns about the records, it can easily request that their accuracy be validated by the shipboard sensors.

Following these explanations, it can be said that the description as to the capability of the autonomous ship to record the sea and weather conditions might also be made a part of the ship’s description under NYPE 2015, so that the charterer might not need to worry about the records and the WRC arrangement indicated under Clause 12(d). Further consideration might also be needed as to Clause 12(e) of NYPE 2015. The clause indicates that where the ship fails to perform as agreed in the charter, the parties may refer the dispute to “an independent expert or alternative weather service by mutual agreement.” These experts resolve disputes over underperformance claims arising from a discrepancy between the parties’ evidence. This clause might be changed for autonomous ships, to state that data collected through the ships will prevail.

2.4.2 The Shipowner’s Obligation to Provide a Seaworthy Ship

Under time charters, for the delivery of the ship to be valid and for the charter period to commence, it is important that the ship meets certain delivery requirements relating to delivery time, the place of delivery, and the condition of the ship, which are commonly set out in the charterparty. These requirements are deemed to be conditions precedent to delivery unless the terms of the contract provide otherwise. Therefore, if the ship does not comply with these requirements, the charterer may reject the ship (Coghlin et al. Reference Coghlin, Baker, Kenny, Kimball and Belknap2014, para. 8.3). In the context of the operation of autonomous ships, a consideration of the requirements stipulated in time charter forms indicates that those relating to the time and place of delivery may not require reevaluation. However, a more thorough analysis is necessary regarding the requirements concerning the ship’s condition.

Standard time charter forms contain terms that specify how the condition of the ship should be at the time of delivery.Footnote ⁶⁹ Most of the forms require the ship to be seaworthy at the time of delivery.Footnote ⁷⁰ The shipowner’s seaworthiness obligation under NYPE 2015 is defined in Clause 2(b):

Even if the charter does not refer to any express seaworthiness requirement in this regard, this is implied by law.Footnote ⁷¹ The classic test for seaworthiness at common law is expressed by Scrutton L.J. as follows:

If this question is answered in the affirmative in a particular case, then it can be said that the ship is unseaworthy. The concept of seaworthiness, in relation to traditional ships, enters the analysis mainly under three groups: (1) physical defects of the ship and inadequate and insufficient equipment; (2) inadequacy and insufficiency of the crew; and (3) inadequacy of documentation.Footnote ⁷³ The legal principles in this regard are well established, and a substantial body of literature has already been written on this topic (Coghlin et al. Reference Coghlin, Baker, Kenny, Kimball and Belknap2014, ch. 8; Girvin Reference Girvin2011, 383–399; Kassem Reference Kassem2006; Zhan and Zang Reference Zhan and Zhang2023). The shipowner’s obligation to provide a seaworthy ship will continue, but these three groups should be reinterpreted and subject to potential change for autonomous ships, considered below.Footnote ⁷⁴

2.4.3 The Shipowner’s and Master’s Obligation to Comply with the Charterer’s Orders

In the context of a time charter, the shipowner, in consideration of payment of hire, grants the charterer the right to exploit the ship’s earning capacityFootnote ⁸⁸ and undertakes to provide a specified charter service. Consequently, the time charterer is entitled to provide instructions to the master about how the vessel is used. Almost all time charter forms contain a provision commonly referred to as an employment clause, which delineates the scope of the charterer’s authority to issue directives to the master concerning the employment of the ship.Footnote ⁸⁹ For example, Clause 8(a) of NYPE 2015 provides: “The Master … (although appointed by the Owners) shall be under the orders and directions of the Charterers as regards employment and agency.” In the absence of an express stipulation in this regard, the obligation is implied because it is necessary to ensure that the contract efficiently works. It should be noted that the master is not obliged to comply with every single order given by the charterer. Under a time charter, the master is obliged to comply with the charterer’s instructions as to employment, but does not have such an obligation if the instructions affect the navigation or safety of the ship. This is a well-established principle of English jurisprudence.Footnote ⁹⁰ The distinctions between employment and navigation are articulated by Lord Hobhouse in The “Hill Harmony” case: “‘Employment’ embraces the economic aspect – the exploitation of the earning potential of the vessel. ‘Navigation’ embraces matters of seamanship” (at 159). Since the charterer’s order regarding the selection of the ports of loading and discharge and orders relating to loading, carriage, and delivery of the cargo embrace an economic aspect, in an ordinary case these must be regarded as lawful employment orders (unless they affect the navigation or safety of the ship, cargo, or crew). Therefore, the master has a duty to follow these orders. Orders that affect the navigation and safety of the ship are left outside the charterer’s right to use, as these are “matters falling within the specialized professional maritime expertise of the master” (at 152).

When this obligation of the master is considered in relation to an autonomous ship, the first question is who (in the absence of any master on board the ship) will be responsible for analyzing whether the order given by the charterer has the status of an employment order that the master is obliged to follow, or affects the navigation and safety of the ship, which the master may disregard. This function can be performed by remote operators.Footnote ⁹¹ It has also been suggested that the word “master” in the abovementioned employment clause of NYPE 2015 should be replaced by “the owners and/or the person or persons operating the vessel” (Baughen Reference Baughen, Soyer and Tettenborn2021, 95). Considering that an autonomous ship is operated under a dynamic level of autonomy,Footnote ⁹² such an approach seems sensible. In a particular case, even if the ship is operated fully autonomously during service at the time of charter, when the order is given by the charterer, the level of autonomy can be reduced to remote operation. The controller can evaluate orders given in terms of safety and navigation and then decide how to proceed. In other words, here the remote operator has started to act in the capacity of the master in assessing the orders,Footnote ⁹³ and the remote operator’s “professional maritime expertise” is brought into the assessment. In such a case, there might be no need for further analysis of this obligation, as the established rules relating to the master’s assessment of the employment orders may also apply to remote operators. For example, when the order is given, it is not expected that the master is obliged to comply with every order immediately. There may be some orders that require further consideration. In such cases, the master has the right to evaluate the order for a reasonable period of time.Footnote ⁹⁴ For an autonomous ship, such time is also provided to the remote operator. Similarly, in terms of a traditional ship, if compliance with the order in question does not fall within the scope of the charter, such as the charterer’s last voyage order, causing delay, the master has the right to refuse to comply with it and to ask the charterer for a new order.Footnote ⁹⁵ This can also be applied in a similar way to autonomous ships. In such cases, the remote operator can ask the charterer for a new order.

However, considering that the remote operator is expected to be responsible for the operation and monitoring of more than one ship, it might not be practical that the autonomy level during a particular voyage is reduced and then every single order of the charterer is evaluated by the remote operator. Therefore, it might be useful to leave some place for the independent decision-making ability of the ship under full autonomy. It will be evaluated below to what extent the autonomous ship may itself evaluate the charterer’s orders under three groups without human support.

The first group comprises employment orders, with which the master is under an obligation to comply. Unless it affects the safety and navigation of the ship, it is expected that the autonomous system also complies with these orders. It is unclear whether the system is capable of making such an assessment. At this point, it might be useful to mention how these systems are to be operated.

AI applications mimic human cognitive and decision-making processes, achieving navigation safety on par with, if not exceeding, vessels manned by humans. In doing so, human-induced risks may be eliminated altogether. AI systems do not merely execute preprogrammed instructions, although it is impossible to program a system to account for every potential problem and solution. Rather, these systems must be capable of adapting to dynamic conditions, much like the crew members they aim to replace, and be able to resolve unforeseen issues (Porathe et al. Reference Porathe, Home, Rødseth, Fjørtoft, Johnsen, Haugen, Barros, Gulijk, Kongsvik and Vinnem2018, 422). As such, it is highly probable that machine-learning methods will be predominantly employed in the management of these ships (Ringbom and Collin Reference Ringbom, Collin, Ringbom, Røsæg and Solvang2021, 10; Vartdal et al. Reference Vartdal, Skjong and St. Clair2018, 10). During machine-learning processes involving technology that enables computers to perform complex tasks by learning from examples, data, and experience (Royal Society 2017, 16), humans can play an active role, either by categorizing or labeling the data for the machine to analyze or, in a more limited capacity, by correcting the machine’s erroneous predictions or selections (Wright Reference Wright2020, 54–55).Footnote ⁹⁶

In some cases, the ship itself may be capable of evaluating the employment status of the charterer’s order. For example, if the charterer’s order is to depart from the port at a certain time, as that time approaches, the ship itself can evaluate the order in terms of safety, using data on weather and sea conditions collected by the ship’s sensors. The system can then determine the most appropriate response to these conditions. Similarly, when the charterer orders the ship to slow steam or to proceed with full speed, the ship can evaluate whether it is required to comply or whether the order should be refused for safety or navigational reasons in accordance with the algorithm, engine capacity of the ship, and charterparty performance warranties. Recall, however, that time charters, by their very nature, rely on continuous adaptability to accommodate instructions from the charterer, and each order introduces potential unanticipated occurrences. Therefore, if the order given is the one that the ship itself cannot evaluate and make a decision about within its algorithm, human support might be needed. In such a case, the autonomy level can be reduced swiftly to remote operation, and the remote operator’s approval, or final decision, can proceed based on the charterer’s orders.

The second group comprises the orders that the master has a right to refuse. These are the orders that fall outside the scope of the agreed charter. If the effect of the employment order given by the charterers is to require the master to do something outside that scope of the agreed charter, the order requires the shipowner to provide services that it has not agreed to provide at the beginning. The master has a right to refuse this order. When the vessel is operated fully autonomously, it may be possible for the ship to evaluate these orders, taking into account the geographical and time restrictions of the charterparty. When the order is given, the autonomous system of the ship without any human support can consider algorithms related to agreed limitations and then refuse to comply with the order on the ground that it is outside the scope of the agreed charter terms. For example, if the last voyage order exceeds the agreed charter period given by the charterer, the system can calculate the length of the proposed voyage and then refuse it. Similarly, if the order causes the ship to proceed to a port outside the agreed geographical limits, the system can refuse to comply with the order. The system can correctly analyze the situation and refuse these orders on the ground that they are outside the scope of the charter. The master has a right to refuse the second group of orders. However, the master may also prefer to comply with this group of orders even though they exceed the limits of the charter. If an autonomous ship makes a decision that the shipowner does not like, a remote operator could come into an active position to overrule it.

Lastly, the master is under an obligation to refuse the third group of orders. These are orders that cause illegal and fraudulent acts (Williams Reference Williams2013, 148).Footnote ⁹⁷ The common example relates to a bill of lading. For example, after the cargo has been loaded on board the ship, the charterer may ask the master to sign the date of the bill earlier than the actual date. Such an order is particularly likely to be given if the charterer has entered into a sales contract under which it, as the seller, is obliged to complete the loading of the cargo on the specified date. In order not to breach its obligation under the contract of sale, the terms insist that the master signs the bill of lading as an ante-date bill of lading. Similarly, the charterer may also order the master to sign the date of charter later than the actual date (that is, postdate the bill of lading). In some cases, the charterer might also order the master to record the goods as being in apparent good order and condition, although the goods are not in such a state.Footnote ⁹⁸ Since the purpose of such orders by the charterer is to defraud the consignee, the master is obliged to refuse them. The master is also under an obligation to refuse the charterer’s order to deliver the cargo to a person who does not have a right to receive it, unless there is an express term to the contrary in this regard.Footnote ⁹⁹

Following this information, the question is whether the autonomous system can evaluate this type of order. Even if the ship is eligible to perform loading and discharging operations autonomously, given that a paper bill of lading is still required, since there will be no crew on board the ship, the evaluation of these types of orders can only be possible if there is involvement of a shipowner’s land-based personnel. This is not something that can be done by the autonomous system or by the remote operator from another center. In terms of autonomous ship operations, if the electronic bill of lading is an issue, the author believes that land-based personnel support is still needed, but signing the bill of lading as ante or postdate can be prevented by the autonomous system.

Continuous reference to human oversight for routine operations would undermine the intended efficiency of autonomous systems. Even if the autonomous system has a role in the assessment of the charterer’s order, it cannot be fully delegated to technology. An autonomous ship should do what it can to the extent of its capabilities, but if it encounters unforeseen risks or conditions that fall outside the scope of its programming and machine-learning process, it becomes essential to rely on the approvals and decisions of remote operators as to the status of the charterer’s order. Following this, it is now clear that Clause 8(a), the employment clause, of NYPE 2015 should be redrafted and the obligation to comply with the charterer’s order should be imposed on “the owners and/or the person or persons operating the ship and/or the autonomous system itself.”

Where the time charterer’s orders have been assessed by the ship’s independent decision-making ability, in some cases legal issues may also arise as to the indemnity right of the shipowner. Indemnity is normally implied when the shipowner suffers loss or liability because of complying with the charterer’s orders, unless the shipowner has agreed in the charter to bear the loss, damage, or liability in question (Coghlin et al. Reference Coghlin, Baker, Kenny, Kimball and Belknap2014, para. 19.32), or unless the order in question is one that the master is under an obligation to refuse (para. 19.19). The principle of causation is integral to indemnity claims. For shipowners to recover under implied indemnities, there must be an unbroken chain of causation between the instructions of the charterers and the loss suffered.Footnote ¹⁰⁰ The advent of autonomous systems seems to challenge this direct causation paradigm. For example, if an autonomous ship grounds at an ostensibly safe port, as in The “Erechthion” case,Footnote ¹⁰¹ causation might not be straightforward. Was the grounding due to defective programming, erroneous input data, the inherent unsafety of the port, or the charterer’s order? The court’s finding in The “Erechthion” – that the grounding was proximately caused by compliance with charterers’ orders – could be difficult to transpose onto autonomous scenarios. Moreover, The “Island Archon”Footnote ¹⁰² demonstrates that liability flows from charterers’ orders when risks are not contractually assumed by the owners. It was held by the Court of Appeal that an indemnity was to be implied, since the risk of incurring liability for dubious cargo claims was one that, under the charter, the owners had not agreed to bear. An analogous argument could arise with autonomous ships where programming errors exacerbate port risks. The legal system must address whether such errors constitute an extension of the charterers’ orders or an independent navigation failure.

2.4.4 The Charterer’s Obligation to Use the Ship between the Safe Ports

The advent of autonomous ships also plants a flag on a reevaluation of traditional principles, such as the safe port obligation. The charterer’s obligation to send ships only to ports that can accommodate them safely is one of the key aspects of a time charterparty. This obligation is expressly stipulated under standard charter forms.Footnote ¹⁰³ Clause 1(b) of NYPE 2015 states: “The Vessel shall be employed in such lawful trades between safe ports and safe places within the following trading limits [text to be entered here] as the Charterers shall direct.” The existence of such a provision primarily requires the charterer to direct the ship to a prospectively safe port.Footnote ¹⁰⁴ This kind of provision is also construed to mean that if the port becomes unsafe later due to a change of circumstances while the ship is still en route, the charterer then has a secondary obligation to cancel its previous order and provide a new, prospectively safe port for the vessel.Footnote ¹⁰⁵ The reason behind imposing a secondary obligation on the charterer regarding unsafe ports is mainly to protect the shipowner and also the crew and any third-party interests.

The automation and digitization of ships and ports redefines what constitutes the well-established general principle of a “safe port” and, particularly, the meaning of “safety.” Therefore, the principle in question, rooted in protecting vessel integrity, requires recalibration to address the uniqueness posed by autonomous ships. The notion is intricately linked with the master’s responsibility to navigate the ship safely and the charterer’s obligation to nominate safe ports. This recalibration can be made through reference to the earlier decisions embodied in case law, as stated in The “Mary Lou”: points of law could arise that were not covered by the general principles, and it might then be helpful to consult earlier authorities, albeit that the conclusion of the court on such points of law should be consistent with those general principles.Footnote ¹⁰⁶ The seminal definition of a safe port was articulated by Sellers L.J. in The “Eastern City”:

The question of whether a port is safe for a particular vessel at a particular time is a subjective one, and the answer depends on the particular circumstances of each case. As English case law developed over time, various modalities emerged to define the concept. These include the identification of dangers as either temporary or persistent, the establishment of monitoring and reporting systems, the consideration of delays and inordinate duration, the evaluation of local warnings and departure systems, the adequacy of warning systems, the analysis of deficient berthing and mooring facilities, the evaluation of navigational aids, and the assessment of safety in departure. Considering these modalities, this analysis will first focus on the meaning of safe ports. Such an analysis is necessary to decide whether the charterer is in breach of its obligation to nominate a safe port under the time charter in the case of autonomous ships. The charterer’s duty to nominate a safe port must extend beyond mere physical hazards to encompass the burgeoning risks associated with cybersecurity. In this regard, digital resilience is an integral component of the “safe port” doctrine (Shazi Reference Shazi2024, 3, 14).

2.4.5 The Charterer’s Obligation to Payment of Hire-Off Hire

The payment of hire by the charterer in consideration of the use of the ship is another obligation for consideration in relation to the autonomous ship. This obligation is expressly stipulated under Clause 17 of NYPE 2015. In the case of autonomous ships operating under time charter, it is important to acknowledge that the charterer’s obligation remains unchanged. The expectation is that the remote operation services are arranged by the shipowner. In most cases, these services will be provided by third-party shipping management companies upon agreement with the shipowner, rather than the shipowner setting up its own center. Since the time charterer will be outside the contractual relationship between the shipowner and the center, the charterer should not be exposed to any further obligation as to payment made in consideration of remote and autonomous operation of the ship. This responsibility should remain with the shipowner. The author suggests that, in the context of autonomous ships engaged in time charter operations, the discussion surrounding payment of hire is likely to be predominantly focused on the concept of off-hire. An off-hire clause is a common feature of time charters.Footnote ¹²⁶ While the exact wording of these clauses varies among charters, it serves to release the charterer from its obligation to pay the hire for a period during which the ship cannot be used due to circumstances beyond its control.Footnote ¹²⁷ Clause 17 of NYPE 2015 contains two main parts. The latter part of the clause addresses the issue of loss of time resulting from a decrease in the ship’s speed. This constitutes an alternative avenue for loss recovery for the charterer in cases of underperformance of the ship.Footnote ¹²⁸ The section requiring further consideration in terms of an autonomous ship is the initial part of the clause, which refers to particular events:

There are two potentially problematic issues surrounding the off-hire clause in NYPE 2015 regarding autonomous ships in time charters. The first is the impact of the text “preventing the full working of the ship.” Although the clause quoted above is silent as to whether the charterparty must be completely interrupted or whether the partial loss of the service provided in a particular case is sufficient to accept that the full operation of the vessel is prevented, the NYPE off-hire clause is a net loss of time clause.Footnote ¹²⁹ Therefore, a partial loss of the charterparty service is also accepted as sufficient for the charterer to benefit from the off-hire clause. Following this, if an autonomous ship operation is prevented partially, for example, it can still be said that the charterer benefits from the off-hire clause.

The second problematic issue is related to events that trigger the off-hire clause. It is accepted that the vessel will be off-hired if these specified events prevent the full operation of the ship. Some of the italicized text identifies events that may still have a role in terms of autonomous ship operations. For example, issues pertaining to the radar system or sensor of the autonomous ship can be interpreted under the “machinery or equipment” group. Although the first part of the events responds, to some extent, to the need for autonomous ship operations, the problems that may arise in relation to the remote operation center, such as a physical deficiency of the center preventing the full operation of the ship, do not fall under any of the events listed. This is normal because autonomous ship operations were not considered when the clause was drafted. It is also important to explore whether at least the second part of the events expressed in the clause under the wording of “by any other similar cause” prevent the full working of the vessel and its lash-up to the operation center.

At first sight, the text appears to cover all types of events that prevent the full operation of the vessel. It may be suggested that all types of events that occur during the operation of an autonomous ship and prevent the full working of the ship could fall under this group. Following this, during the operation of an autonomous ship, if there are any problems related to the remote operation center or personnel in the center, the charterer could benefit from the off-hire provision under this category. However, due to the existence of the word “similar,” the text must be regarded as referring to the same types of causes as those previously mentioned in Clause 17.Footnote ¹³⁰ The term therefore will only apply to other similar causes listed in the first part of the clause. When the events listed in the first part above are considered in terms of autonomous ship operations, none of them appear related to events that might arise at a remote operation center. Therefore, it is difficult to include the loss of time in relation to the operation center and personnel in the center in the “any other similar cause” wording. If the word “whatsoever” is inserted and the wording in the NYPE 2015 off-hire clause is redrafted as “any other cause whatsoever,” there will be no need for the causes to be the same as those in the list to trigger the off-hire clause.Footnote ¹³¹ This approach underscores the dispositive effect of the word “whatsoever.”Footnote ¹³² In such cases, problems relating to remote operation centers or personnel at the centers can fall within the second part of the events and the charterer can gain the benefit of the off-hire clause. However, inclusion of the phrase “whatsoever” will widen the scope of the off-hire clause considerably, so it will be challenging to convince the shipowner to accept this new text. Instead, the easier solution would be that, considering the nature of autonomous ship operation, the list of events given in the initial part of the clause is expanded to include new events sui generis to the remote operation center, such as a breakdown of the connection system, software issues, and data transmission problems. It should be noted that although these problems arise, in some cases this might not prevent the full operation of the ship and loss of time may not arise. For example, if the ship is operated remotely and its control and operation are lost due to a failure in data transmission, and the ship’s autonomous operation feature comes into play (as a result of the dynamic autonomy of the ship) and the ship is still able to operate on its own, there will be no loss of time, so there will be no off-hire period. In addition, it is possible that the service provided in the operation center is disturbed due to fire, flood, earthquake, or other natural events, so that the full working of the ship is prevented. Such events might also be incorporated into the clause. As given above, the off-hire clause also refers to particular events related to the crew – such as deficiency, default, or labor strike – and those types of events might also arise at the center. Such a situation is, of course, more manageable than at sea, but might still cause loss of time for the charterer. Therefore, a deficiency, default, or labor strike of personnel in the center should be included in the clause.Footnote ¹³³

Given that the occurrence of some types of events, such as communications or software problems, may regularly disturb the operations of autonomous ships, the incorporation into the clause of a provision for de minimis events is sensible. Such a provision makes the activation of an off-hire clause conditional on the minimum period required by the clause to have been lost due to events specified in the clause.Footnote ¹³⁴ For example, according to the de minimis provision in BPTime 3, the ship is not considered to be off-hire in a case in which the loss of time is less than three hours in any calendar month. Under that clause, the charterer is deemed to have calculated the time lost due to an off-hire event incorrectly if it takes into account a loss of time less than three hours. The addition of such a provision in NYPE 2015 could prevent the expected but common problems related to software and connection problems from causing legal issues between the parties if they occur for only a short period of time.