Method of investigating collisions will not change fundamentally with autonomous shipping

In the third of HFW’s six-part series on a fictional collision between an autonomous cargo ship and a container ship in Felixstowe harbour, Associate and mariner Jonathan Goulding, along with partner and master mariner Jim Cashman, said that the principles of investigation for accident investigation boards and for the courts are not likely to change greatly with the introduction of autonomous shipping. The MAIB (for example) would continue to investigate accidents in order to find out why they happened and to attempt to ensure that they are not repeated. The courts (and, therefore, the lawyers) would also have a view to identifying breaches of regulations and whether those breaches contributed to the incident. Neither Goulding nor Cashman could recall offhand a case where both vessels were in motion that there was an allocation of 100% liability to one side and 0% liability to the other.

However, what would change with the introduction of autonomous shipping would be the nature of the investigation and the type of experts called upon to identify fault.

Goulding observed that with the introduction of MASS the place where fault lay would also potentially change. Would it be the ship’s captain (and how would one decide who was the captain on an autonomous vessel?) or would it be the software programmer responsible for an algorithm that functioned in an unintended way, or would it be the person responsible for a patch that should have been installed, but wasn’t, or was installed incorrectly?

Goulding pointed out that autonomy was really a stage in the evolution of shipping rather than a revolution. The introduction of container ships had led to an increasing automation of loading and unloading, and the number of people on a ship had been falling for decades.  On the fact-finding side for lawyers, VDRs and AIS (despite AIS, or indeed VHF radio, not being mentioned in the laws of the land in England) plus radar makes the reconstruction of ship routes and the events leading up to any incident far easier than it used to be. Goulding said that autonomous shipping should make such reconstructions easier still, given the higher amount of data available.

Referring to the fictional scenario in Felixstowe, Goulding said that in this case an AL3 (Lloyd’s classification – Decisions and actions are performed autonomously with human supervision and authorization) lost control due to a cyber event, a hack. The principles of inter-ship liability and the method to apportion fault and liability were unlikely to change in such a case, Goulding said. He hypothesized that a future case might compare a ship that loses control due to a hack might be compared with a ship that becomes “not under command” or one where control is lost due to an old-fashioned mechanical failure. Instead of an engineer seeking to identify the cause of the engine failure, a software engineer will be looking at the code to identify the cause of the software failure. “Will our position change? My initial view is ‘no'” said Goulding.

In summary, HFW said that it was unclear how the investigation of marine casualties would change with the introduction of MASS vessels. The industry had been undergoing relatively rapid changes in recent decades and, though a little behind the curve, investigation and claims had adjusted to the changing the nature of the industry. Although there might be some additional considerations, any future changes to allow for this technology were likely to be managed in a similar manner.

Cashman said in response to a question that lawyers would still be there to apportion fault because the shipping industry had seen many opportunities to establish a knock-for-knock system, and had never done so. When shippers suffered a loss, it was human instinct try to pass on that loss.

Looking even further ahead to the time when there might be international trans-ocean autonomous shipping, Cashman observed that autonomous was not the same as unmanned. There could be two people on large vessels only there in case something went wrong. However, Cashman warned that there would then be pressure on the shipowners, after a year or two of these people doing nothing and no incidents occurring, to get rid of one of the remaining crew of two, and finally to get rid of the last one. Neither would they necessarily be seafarers in the first place.

Referring to the pair of people on board an autonomous ship apparently doing nothing, Goulding concluded by saying that two key elements for autonomy in the future would be redundancy and reliability. How reliable were the systems and what was the degree of back-up when something went wrong? A lack of redundancy could be as significant as a lack of reliability when it came to apportioning fault.

*In yesterday’s IMN it was reported that Jonathan Goulding said the cyber-exclusion clause 380 would “fade away” by 2020. This was a typo. In fact he said that it would fade away by 2030. Apologies.