Trieste - Among the laws of maritime law that seem to hinder the advent of unmanned navigation in a more sensitive way, the International Rules for preventing collisions at sea (ColReg) have particular significance.
Prepared by the IMO and implemented by the repeatedly amended London Convention of 1972, they establish the rules of conduct and maneuver to prevent collisions at sea and to establish fault in cases of collision.
Taking the general provisions relating to their field of application in mind, addressing themselves to “every vessel” the Rules are neutral in relation to the existence of the crew on board.
The proof is that since the examination of Rule 2 - that which appears to be of greater importance among the ColRegs - the first critical points emerge. This provision establishes the prevalence of good seafaring practice with respect to strict observance of the provisions established by the Rules on governance and maneuver, explicitly clarifying the principle that in certain circumstances it may be mandatory to depart from the Rules.
Rule 2 requires that the exercise of human discernment and the carrying out of the decision-making process in the nautical management be contextual, especially in those situations where it is necessary to decide whether it is necessary to perform a maneuver prescribed by the ColRegs or, on the contrary, it is instead required a completely different one.
In principle, this evaluation can be carried out remotely, to the extent that adequate data transmission and communication technology is available. And this seems to be in fact the most widespread conviction: the operators of the ground control station, assuming responsibility for the management of the autonomous ships with remote control, will be called to fulfill the rules of conduct imposed by the ColRegs, or to depart in accordance with the Rule in comment, as invested with command functions.
In this perspective the obligation, established by Rule 6, concerning the determination of the safety speed, will pass from the figure of the Commander to that of the manager of the remote control station.
Therefore the requirement established by the Rule 2 would probably seem to be satisfied both in the case of navigation with remote control, and in that carried out with autonomous ships but subject to permanent human surveillance and of which it is possible to take remote control.
On the other hand, the rule in question does not appear to be capable of being applied to autonomous navigation carried out without human supervision in the current text and its reformulation has therefore been suggested.
Coming to the rules of government and maneuver, Rule 5 seems to be one of the main obstacles to the use of autonomous ships, since its content (and in particular the reference to the lookout service) would seem to suggest that, for the purposes of its fulfillment, human persons must necessarily be present on board.
In reality, the main functions of the lookout are to collect information about what surrounds the ship and to convey this information to those who can evaluate it and make the decisions that are suggested by the related context. It is therefore a type of activity in which autonomous technology, insensitive to fatigue, decreased attention or situational distraction, excels compared to humans.
Nevertheless, the reference to “sight and hearing” seems to impose a human contribution in the detection and assessment of the situation and the risk of collision.
With regard to remote control ships, the problem is therefore not particularly marked. The autonomous vessels of more recent conception are in fact equipped with sophisticated audio and video sensors, used to transfer the representation of what surrounds them to the person in charge of the remote control, which therefore will carry out with regard to the duty to exercise adequate supervision, the tasks currently burdening the commander. This presumably satisfies Rule 5, preserving the role of human input in the evaluation process, in the sense that neither the use of electronic instruments (similarly to what is already happening today with regard to radars) nor the fact that orientation occurs from ground control station means that their combination betrays the spirit of the provision in comment.
The last aspect to be considered with regard to the rules of governance and maneuvering pertains to the Rule 18 concerning the “Responsibility between ships”. In particular, it assigns navigation priority to the ship “which does not govern”, defined by rule 3 (f) as that “ship which for exceptional circumstances is unable to maneuver as required by these rules and is therefore not able to leave free the route to another ship.”
This rule usually refers to ships in difficulty, for example due to engine failure. In the unmanned ship sector the condition of “ship that does not govern” first of all evokes the risk of the failure of the ship’s communication system.
The solution to problems of this nature could perhaps be seen in the use of appropriate signaling systems to inform those navigating near the damage. However, the reference of the norm to “exceptional circumstances” seems to limit its scope to different profiles with respect to the ordinary mode of operation of the ship, which could perhaps make it difficult to subsume the autonomous management of the ship in the field of operation of the provision in question.
In summary, while the advent of navigation with remote control and autonomous navigation under human control do not seem to impose a radical modification of the ColRegs so that they can regulate the phenomenon, just as it does not seem to be able to be said with regard to fully autonomous navigation, in the state incompatible with the their current formulation.
As mentioned, the ColRegs detect to establish the fault in the event of a collision and it is precisely in relation to the responsibility for collisions that autonomous navigation poses the greatest criticalities.
With regard to autonomous navigation, the identification of new liability rules arises in terms of legislative policy, in which essentially three scenarios are envisaged: (i) objective liability of the shipowner, accompanied by the obligation of adequate risk insurance and by a direct action by the injured party; (ii) objective liability of autonomous vessels; (iii) liability for fault of the software and hardware developers, system engineers and stand-alone shipbuilders.
This same approach can be seen in the publication that deals with the issue of channeling responsibility for autonomous driving cars, where there is no lack of those who maintain that, given the capacity for autonomous thinking of artificial intelligence, to the point of allowing it to depart from the rules imposed on it from the manufacturer, it is possible to assign to the autonomous vehicles legal personality.
The question, therefore, is not one of technique, but of legislative policy: the reference to the fault of the ship contained in art. 3 Conv. Brussels 1910 and the implicit reference to the common law doctrine of the personification of the ship and to its possible corollaries (first of all the actio in rem), already explains that in technical terms attributing to the ship juridical personality through the fictio is a road, although distant from the civil categories and from the continental sensibility, which can be traveled.
The issue on which we need to reflect is whether, as much as possible, such a step is desirable.
It is therefore necessary to ask whether it makes sense to imagine imputation of responsibility to a ship without this corresponding to a full autonomy, not already and not so much in terms of navigation, but rather with regard to the channeling to the ship of every interest and activity relating to the operation of the shipping company.
This is not reflected in today’s social and economic reality and the answer should therefore be negative, appearing in the most appropriate state to channel the responsibility towards the shipowner who takes the exercise of the ship.
* Senior Partner, Zunarelli Studio Legale Associato