Artificial intelligence and the advent of automated ships will pose difficult questions in determining liability under the Hague Rules for maritime casualties, the 2022–2023 chairman of the Association of Average Adjusters has cautioned.
Sir Nigel Teare raised the concern as he suggested that a recent case, in which the Supreme Court confirmed that a defective passage plan will render a vessel unseaworthy, will be scrutinised in the context of technological developments.
Sir Nigel chose for his annual chairman’s address to the association in London the topic of Seaworthiness, Negligent Navigation and Safer Ships, reviewing the implications of the general average case that he tried at Admiralty Court level in 2019, relating to the containership CMA CGM Libra.
The laden vessel left the buoyed fairway and grounded as she was leaving Xiamen, one of the largest ports in China, in May 2011, necessitating a costly salvage operation. Most of the cargo interests accepted that the cause of the casualty was negligent navigation and paid their contribution to the general average, but a small minority refused to do so. The shipowner failed in proceedings in the Admiralty Court to recover general average sums from that minority, and its decision was upheld by the Court of Appeal and the Supreme Court.
In his address to the association, Sir Nigel extended the question of passage planning to its potential application to vessels controlled by operators ashore or by computers on board—ships known as Maritime Autonomous Surface Ships, or MASS for short. “They are not yet common,” he said, “but, with commendable foresight, MASS are being closely monitored by the International Maritime Organisation.” The IMO is exploring how far regulatory regimes such as Solas and the Collision Regulations can apply to autonomous ships. Its plan is to have a non-mandatory code for that type of ship by 2024, with a mandatory code in force by January 2028.
There was no reason why MASS should not have a passage plan to be seaworthy for their voyage, said Sir Nigel. “Such plans may in the future be produced by a computer. What if the reason that a passage plan is defective lies in a defect in the software purchased by the owner?”
He noted that Sir Richard Aikens, another distinguished judge, has suggested that in such a case, an owner would be able to say that he had committed no breach of due diligence because the defect came about when the ship was not under his control, or, as it was put in one case, ‘in his orbit.’ “That, if I may say so, appears to be correct,” remarked Sir Nigel. “The need for the vessel to be in the orbit of the owner before the owner is subject to the due diligence obligation in the Hague Rules [the protocol of 1924 on the allocation of risk between shipowners and cargo interests] was recognised by Lord Hamblen in the Supreme Court in the CMA CGM Libra case.
“Thus, the focus will be on the question whether the owner, once he was in possession of the software, should have appreciated, by careful and skilled monitoring of the software, that it was not in a fit condition for its purpose. If his monitoring was negligent, then there would have been a failure to exercise due diligence. I suspect that it will be difficult to establish negligence of this nature.
“Where the master on board or operator ashore acts negligently when commanding the vessel, that would amount to negligent navigation. But what if the error is committed by a computer? If such an error is the result of a defect in the software, the automated vessel would presumably be regarded as unseaworthy. Again, there would be no breach of due diligence by the owner unless the owner could and should have detected the error before the commencement of the voyage. But if the error is the result of an error by the artificial intelligence of the computer, then that might well be regarded as negligent navigation by the computer, just as if it had been an error by the officer of the watch.
“Thus,” continued Sir Nigel, “one can envisage expert evidence from software engineers as to the nature of the defect and as to whether the defect could and should reasonably have been discovered by the owner or whether it was simply a mistake by the artificial intelligence of the computer.
“At present, expert evidence in maritime cases is given by former masters and marine engineers. But in the future, as and when MASS suffers a casualty, the important experts may well be software engineers. I would find that a depressing prospect, but more youthful minds may disagree,” he observed.
Sir Nigel insisted that the introduction of electronic charts had not eliminated the need for proper passage planning. To be seaworthy, a vessel must still, at the outset of the voyage, have a proper passage plan. That remained essential to safe navigation.
The AAA chairman recalled that the CMA-CMG Libra case had caused disquiet among shipowners and protection and indemnity clubs. It had been suggested that the decision would lead to more cargo claims against carriers and their insurers and to more claims in general being resisted by cargo interests. Since the decision, Sir Nigel has been informed by the P&I clubs that it is now routine to find allegations of poor passage planning in cargo claims. But in considering whether the decision would materially increase the share of the burden of maritime casualties borne by carriers, it was necessary to ask whether any defect in a passage plan would render a vessel unseaworthy.
Sir Nigel doubted that the effect of the Supreme Court decision would be as damaging to shipowner interests as some had suggested. He said that “if the decision in the case does lead to carriers bearing a greater burden of the losses caused by maritime casualties than in the past, that is because the notion of a seaworthy vessel keeps up with and reflects modern safety standards. That is not a cause for regret. Seaworthiness is the handmaiden of beneficial changes in ship management designed to promote safety at sea.”
He added: “It is difficult, I think, to contemplate a case where a defect in the passage plan does not render the vessel unseaworthy.” However, a note of caution should be sounded. Lord Hamblen had said that in any future case, it would be necessary to prove that the defect in the plan was ‘sufficiently serious’ to render the vessel unseaworthy.
Opinions as to what constitutes proper passage planning might reasonably and legitimately differ. This would have to be borne in mind when deciding whether a sufficiently serious defect had been established. The reality was that the scope of the seaworthiness duty was not fixed in stone but could adapt to and encompass changes in the practise of shipping.
Some in the industry had suggested that the decision in the CMA CGM Libra case was now out of date in the world of electronic charts. “I beg to differ. The advent of electronic charts has not eliminated the need for proper passage planning. To be seaworthy, a vessel must still, at the outset of her voyage, have a proper passage plan. Such a plan remains of essential importance for safe navigation.
“What will become important will be the training of navigational officers in the art of passage planning when using electronic charts. The UK Government’s Marine Accident Investigation Branch has urged navigating officers not to rely solely on the data embedded in a computer-based system but to consult all sailing directions and notices to mariners just as they would when passage planning on paper charts. There is evidence that not all navigating officers understand the limitations of electronic charts.
“It is therefore possible that the introduction of electronic charts will give rise to more, not fewer, complaints of poor passage planning. In circumstances where the decision of the Supreme Court has so clearly resolved the issues of law, the debate in the future is more likely to concern the adequacy of the plan on the electronic chart, the significance of any defects, and the adequacy of the training of officers to use electronic charts safely,” said Sir Nigel.
