Master of an autonomous ship – legal implications: The introduction of unmanned vessels could have wide-reaching implications for questions of class, seaworthiness and more

23 Янв

Автор настоящей статьи раскрывает пути развития автоматизированных судов, современное состояние и перспективы их дальнейшей эксплуатацию в морской индустрии, причем особое внимание уделяется правовым вопросам, связанным с их эксплуатацией, в частности, с возможными изменениями в международных морских конвенциях.

Является ли автоматизированное (автономное) судно судном в правовом смысле? Соответствуют ли такие суда требованиям конвенций СОЛАС и других, в особенности, KTMC и Конвенции ПДНВ, а также Конвенции о Международных правилах предупреждения столкновения судов в море? Есть ли необходимость во внедрении автоматизированных судов в сферу морского судоходства и торговли? Останутся ли в силе те же международные нормы, применимые для судов, укомплектованных экипажем, или возникнет необходимость создать новые специальные конвенции для автоматизированных судов? Каковы будут требования к капитану судна, либо они вообще будут отсутствовать в связи с ненадобностью капитана на судне? Возможно ли понижение уровня безопасности мореплавания и какова гарантия сохранности груза при морской перевозке посредством автоматизации систем судна? Что случится с судами, укомплектованными экипажем, в связи с внедрением инноваций? На эти и многие другие вопросы пытается найти ответы автор публикации в международном морском праве и английском морском праве.We appear to be on the cusp of a momentous change, with the possible widespread introduction of autonomous ships in the not-too-distant future. But what do we mean by ‘autonomous ship’? There appear to be three distinct stages in the move to fully autonomous vessels:
a. Lean-manned
This would be an intermediate step, with many onboard tasks automated, allowing for substantially reduced manning, with comprehensive support from ashore. Management would have fulltime total awareness of shipboard conditions.
b. Remote-controlled
This is the next step on the way to fully automated vessels. Some developers predict the technolog)’ will be available and possibly cost effective by 2020.
c. Unmanned
This would be the ultimate aim, although it is currently envisaged that such ships might be fully autonomous for the ocean passage, but remotely controlled closer to port or in congested waters.
How far along the road are we?
Some autonomous ships already exist in the military and in some surveying roles, but what we are concerned with here is the widespread introduction of commercial and military autonomous vessels effectively replacing much of the currently manned fleet. Developers are envisaging remote-controlled ships in sendee by 2020 and fully autonomous ships by 2030, although there will, inevitably, be a period when both manned and unmanned ships will be at sea at the same time.
In Norway, on an even more ambitious timetable, the world’s first fully electric autonomous container feeder, Yarn Birkeland,
(see illustration, r) is already well under development. This project envisages the vessel being delivered by 2018 and in fully autonomous operation within Norwegian coastal waters by 2020.
What are the legal issues?
Today’s legislation has been drafted on the assumption that a ship includes a Master and crew on board and in charge of navigation, safety and other operations. These laws and conventions will have to be reviewed to ensure that they are applicable to autonomous ships. Lean-manned ships will probably sit fairly easily within the current legislation, but things are likely to be more problematic with unmanned ships, both remote-controlled and fully autonomous.
We are at a very preliminary stage of considering these issues.
For example, the Maritime Safety Committee (MSC) of the IMO concluded in June 2017 that it would put what it calls maritime autonomous surface ships (MASS) on its agenda. It agreed to carry out a scoping exercise, identifying which IMO regulations preclude unmanned operations, which are not applicable, and which ones will need to be changed. This exercise is expected to last until mid-2020.
Even by 2020 the IMO will only have got as far as identifying relevant legislation. It will not have begun to draft amendments to the regulations, let alone draft any new regulations that might be required.
Is an autonomous ship a ‘ship’ in the eyes of the law?
Different definitions of‘ship’ appear in different statutes and conventions, and it has often fallen to judges to decide if a floating object is, or is not, a ship. However, none of the definitions of‘ship’ requires that the floating object be manned, generally simply requiring that it be used or be capable of being used in navigation. Whether this is by remote control or fully autonomous would not appear to be a problem in terms of the legal definition of a ‘ship’. It appears probable, therefore, that autonomous ships are likely to fall within the legal definition of‘ship’ and so will have to comply with the relevant laws and conventions – although the laws themselves might require some modification.
Certification issues
The principal stipulations regarding the construction and certification of ships are contained in the SOLAS Convention, which requires signatory states to ensure that ships flagged by them comply with minimum safety standards in construction, equipment and operation.
This means that a vast body of rules and statutory provisions, as well as the provisions of SOLAS, are going to have to be pored over and possibly amended to ensure that they make sense when applied to autonomous ships so as to avoid problems arising at certification stage. For example, SOLAS (Chapter II-l) requires every tanker constructed on or after 1 July 1998 to be provided with means for the crew to gain safe access to the bow even in severe weather conditions. Unless this provision is amended then even crewless autonomous ships will have to be provided with a safe means of access, failing which they will not be provided with a cargo ship safety construction certificate.
Given that there are likely to be hundreds of such provisions, an alternative might be for an overarching provision to be agreed that regulations relating solely to crew will be null and void where the vessel is designed to operate without a crew. However, the drafting of such an overarching provision could well prove problematic.
Legal and practical issues might also arise for a ship constructed without any crew facilities if, for example, maintenance riding squads need to be carried on board at times. If the ship has no accommodation, cooking or sanitary facilities, then it will not be in compliance with the MLC, which sets down minimum standards for accommodation. In addition, an autonomous ship would, presumably, not carry any lifesaving appliances or certificate.
These comments are, however, based on traditional ships. Because autonomous ships will be newly designed from the keel up, there is going to be something of a paradigm shift here. Should a riding crew be needed perhaps a self-contained accommodation unit could be lifted on board along with a ‘containerised’ lifeboat and freefall davit.
What about crewing requirements?
The principal stipulations regarding the number and qualifications of the crew to be carried by a vessel are contained in the SOLAS and STCW Conventions.
The STCW Convention sets minimum standards for training, certification and watchkeeping for seafarers on seagoing merchant ships. SOLAS covers manning levels. STCW does, however, require an officer and a lookout on the bridge of a ship at all times save in certain circumstances.
Thus the question arises, will unmanned ships be able to comply with the manning requirements stipulated in SOLAS and STCW?
Regulation 14 in SOLAS Chapter V deals with ships’ manning.
It requires governments to ensure that all ships are sufficiently and efficiently manned from a safety point of view. It also requires governments to issue every ship with an appropriate minimum safe manning document.
The IMO has specified guidelines for flag states when assessing safe manning levels. Many of these will be irrelevant to an unmanned ship. However, the guidelines include ensuring that the ship has the capability to:
• maintain safe navigational, engineering and radio watches and general surveillance of the ship;
• moor and unmoor the ship safely;
• perform operations for the prevention of damage to the marine environment;
• maintain the cleanliness of all accessible spaces to minimise the risk of fire
• ensure safe carriage of cargo during transit;
• inspect and maintain the structural integrity of the ship and also deploy a competent damage control party;
• operate the main propulsion and auxiliary machinery and maintain them in a safe condition.
If the flag state is not satisfied that all these functions can be fulfilled safely by an unmanned autonomous ship it is unlikely to issue a Safe Manning Certificate for a ship with zero manning. Without a Safe Manning Certificate the vessel will be unable to operate.
There are other relevant Regulations in SOLAS.
For example, Regulation 33 in Chapter V obliges a Master to respond to information received from any source that persons are in distress at sea. Under English law, any contravention would be an offence by the Master, punishable by up to two years imprisonment or an unlimited fine, or both. But how will an autonomous ship comply with this? Would the ship even know if it passed a raft full of refugees, for example, or a distressed yacht with no radio but indicating distress using flares?
Arguably, if no-one was on board to see the persons in distress then the autonomous ship would not have received any information that persons were in distress at sea, and so no obligation would arise and no offence would have been committed. But is this an attractive approach?
It remains to be seen how flag states and the IMO will deal with this and whether they will be prepared to issue safe manning certificates to unmanned vessels.
The STCW stipulates the standards of training and certification required by crew members. These aspects will not affect unmanned vessels as such. They might, however, have an impact on the requirements for personnel manning the shore control centre (SCC), who will effectively be in command and on lookout on the bridge of the vessel remotely.
The qualifications, knowledge and experience required by a shore- based SCC operator are likely to be very different from those needed by a seagoing seafarer. For example, will an SCC operator have to be trained in lifeboat operation, firefighting, shipboard maintenance and first aid? Will they be subject to the same hours of work and rest regulations and sobriety tests?
As stated above, STCW does require an officer and a lookout to be on the bridge at all times, with certain exceptions. An autonomous ship will not be capable of satisfying these requirements, unless the electronic lookout in the form of radar and other sensors is legally recognised as equivalent to a human lookout.
Who will be the Master of an autonomous ship?
The STCW Convention defines the ‘Master’ as ‘the person having command of a ship’. Will this be the SCC operator? If the SCC is manned in shifts, will there be a single overall Master of a particular ship, or will it be whoever happens to be on duty within the SCC at the time?
There are legal obligations and severe penalties for their breach attaching to the Master of a ship. Will the SCC operators be prepared to accept these obligations?
Will an autonomous ship be able to comply with the Colregs?
Part A of the Colregs contains general provisions which define ‘vessel’ as including ‘every description of water craft used or capable of being used as a means of transportation on water’, and ‘power-driven vessel’ as ‘any vessel propelled by machinery’.
It appears, therefore, that autonomous ships will ordinarily be ‘power driven vessels’ within the meaning of the Colregs.
Part В contains the Steering and Sailing Rules and is divided into three sections:
Section I – Conduct of vessels in any conditions of visibility (Rules 4-10)
Section II – Conduct of vessels in sight of one another (Rules 11-18)
Section III – Conduct of vessels in restricted visibility (Rule 19).
The rules provide for different actions to be taken by vessels in sight of one another and vessels in restricted visibility and they state that: ‘vessels shall be deemed to be in sight of one another only when one can be observed visually from the other’.
Will this mean that other vessels are never in sight of an autonomous
ship? Some clarification will be required on this aspect. Even if autonomous ships do appear from about 2020 onwards, manned ships will not disappear overnight and vessels such as cruise ships will always need to be manned.
If a human lookout on a manned ship detects that the vessel is in restricted visibility it will be essential that an autonomous ship in the same area is also able to do so, otherwise the two vessels will potentially be obeying different collision avoidance rules.
But how will the autonomous ship know if it is in fog? It appears that autonomous ships will not have a remote operator in control at all times and so effectively no-one will be looking out of the window, or rather through the camera, for much of the time.
Rule 5 in the Colregs provides that a proper lookout shall be maintained at all times by sight and hearing as well as all available means. However, unless the autonomous ship is under remote control, its lookout will be by electronic means, principally radar. This would not appear to comply with Rule 5, which might need to be amended.
Rule 6 provides that: ‘Every vessel shall at all times proceed at a safe speed… In determining a safe speed the following factors shall he among those taken into account:… the state of visibility.’
Again, how is the autonomous ship to determine the visibility?
Rule 6 goes on to provide that, when determining a safe speed, vessels equipped with radar should also take into account:
‘… the effect on radar detection of the sea state, weather and other sources of interference; the possibility that small vessels, ice and other floating objects may not be detected by radar at an adequate range… ’
This raises an interesting point. If the Colregs recognise that radar might not pick up small vessels and ice at an adequate range, what will be the legal position if an autonomous ship collides with, say, a small yacht that its radar did not detect?
The very fact that such limitations are officially recognised, and that autonomous ships will be relying very heavily on radar, means that it would be very difficult for the shipowner to argue that such an incident was unforeseeable. The legal consequences could be severe if, say, there was loss of life on the yacht as a result of the collision.
Rule 18 applies to vessels in sight of one another and lists which categories of vessels are to keep out of the way of other categories. For example:
A power-driven vessel underway shall keep out of the way of:
• a vessel not under command;
• a vessel restricted in her ability to manoeuvre;
• a vessel engaged in fishing;
• a sailing vessel.
Possibly this rule could be rewritten to provide that all vessels are also to keep out of the way of an autonomous ship. However, this is less than satisfactory as it might be difficult to know whether a vessel was an autonomous ship or not and it would require all the world’s seafarers to re-learn the Colregs.
Rule 19 provides that:
‘This Rule applies to vessels not in sight of one another when navigating in or near an area of restricted visibility.’
As noted above, there is the unattractive argument that other vessels are never in sight of an autonomous ship as she has no visual lookout (save possibly when being controlled remotely) and so she could always follow Rule 19. However, the Rule applies only when vessels are navigating in or near an area of restricted visibility, so this second element of the rule would appear to defeat the argument that autonomous ships could always apply Rule 19.
Thus, the question again arises as to how an autonomous ship will be able to determine if she is in or near an area of restricted visibility.
Rule 2 in part A, headed ‘Responsibility’ provides that:
‘Nothing in these Rules shall exonerate any vessel, or the owner, Master or crew thereof, from the consequences of any neglect to comply with these Rules…’
This means that incorrect actions under the Colregs leading to a collision will attract liability, as has always been the case. So, if it could be shown that a collision occurred because an autonomous ship was, say, applying the rules for vessels in sight of one another when in fact she was in restricted visibility, she may well expect to bear the major part of the liability for the collision.
The IMO, flag states and class will need to be satisfied that autonomous ships can successfully comply with the Colregs before they will certify such ships to trade.
Cargo and seaworthiness
Cargo care – ventilation, pumping bilges etc – is likely to lend itself readily to automation, provided the input information is available. After all, ventilation is governed by factors such as internal hold temperature, external air temperature and humidity and time of day. All of this should be easy enough for an autonomous vessel to compute. Bilge pumping can be activated by high-level sensors in the bilge wells.
There will still be a need to clean the holds before loading, and to load, stow and secure cargoes. In addition, cargoes may need to be inspected prior to loading. For example, the moisture content of bulk cargoes may be suspect and need to be assessed.
Who will look at the cargo and reject it or clause the bills of lading if appropriate? Failure to do so may endanger the ship, or leave shipowners exposed to claims for cargo damage for which they are not responsible, but without legal defences that they might otherwise have had.
The Hague and the Hague/Visby Rules provide shipowners with defences to liability’ for cargo damage provided they have exercised due diligence to ensure the vessel is seaworthy before the voyage. This is not an insurmountable problem for an autonomous ship, but it is going to be more difficult for owners.
One of the seaworthiness factors is the competence of the crew. This will translate into the competence of the SCC operatives and it will be necessary for owners to be able to show that they engaged competent individuals for this task. How this will be done remains to be seen, as we do not yet have a legal framework for the qualifications and training of the SCC operatives.
What if communication with the vessel is lost, or there is a software glitch resulting in loss or damage to the cargo? Could this be a seaworthiness issue, in which case the owners might lose their defences to a claim from cargo interests? However, they might then have an indemnity claim against the operator of the autonomous ship, if a third party, or against the software provider.
Shore control and legal liability
Schedule 4 of the UK’s Merchant Shipping (Safety of Navigation) Regulations 2002 imposes penalties for breach of the regulations contained in SOLAS Chapter V. The penalties, of up to two years’ imprisonment or an unlimited fine, apply to the Master and/or the owner of the ship, or both, depending on the breach.
The ‘owner’ is defined as including: ‘Any person or organisation, including the manager or bareboat charterer, who has assumed responsibility for the operation of the ship from the owner.’
This potentially raises some pertinent questions as to whether legal liability would he with the owner of the autonomous ship or those controlling it remotely, if this is done by a third party. We can look forward to some interesting contractual negotiations when these contracts are being drawn up.
The author can be contacted at
Editor’s note: This is an edited version of the original article. The
full article is available on request from the editor.

Автор: David MacLeod, MNI,
Solictor & Consultant, Salvus Law Ltd
Источник: Seaways. – 2017. – December. – P. 12 – 14.