"Any man’s death diminishes me, because I am involved in Mankind;
And therefore never send to know for whom the bell tolls; it tolls for thee."
–John Donne
The recently published Marine Accident Investigation Branch (MAIB) report for 2025 has two cases that directly concern tugs, and they are extremely worrying, not least because they occurred in major ports in the United Kingdom, where safe towing should be practiced as a matter of course.
In the first case, a small tug girded and capsized with the tragic loss of two lives. Among the failures identified by MAIB were: the towing plan was not understood and agreed by all parties; the tug’s gob rope was unable to withstand the force exerted on it; the ship’s speed exceeded port guidelines; and open watertight doors contributed to the tug sinking.
Inadequate training and experience were also cited.
MAIB made a number of recommendations, most of which have either been implemented or are about to be implemented.
Jobs that should have been routine turned deadly not through unexpected turns of events, but through what should have been foreseeable factors.
In the second case, a towing pennant failed and the snapback shattered a wheelhouse window and injured a crewmember. Among the failures identified were: insufficient induction for temporary crew due to commercial pressure; ineffective vessel condition assessment; ineffective towing winch maintenance which led to its contamination and subsequent slippage; failure to identify the pennant as unfit for service; and inability of the wheelhouse windows to withstand towline snapback impact.
Once again most of the MAIB recommendations have been accepted, as you would expect from a responsible tug company. What concerns me, though, is the number of basic mistakes that crept in. Jobs that should have been routine turned deadly not through unexpected turns of events, but through what should have been foreseeable factors – a towing plan that was not understood and agreed by all parties, excessive speed, open watertight doors, insufficient induction training, ineffective maintenance and the rest.
These are things that the people involved should have been aware of, and probably were, but for some reason the issues were not resolved. Both cases remind us of the need for constant vigilance on board, prompt action when things are not right, and support from management to resolve the problems.
If this can happen in established ports with above-average safety and training systems, how much worse could it be in less well-run jurisdictions.
And there are always lessons to be learned, even for those of us who would insist we would never leave a watertight door open during towing. In my case, it was the wheelhouse windows that were my undoing. When I was involved in building tugs, I always made sure the windows were of toughened safety glass, but I simply assumed that classification society rules would ensure the glass could withstand the impact of snapback or the occasional large wave. Apparently not, and I will avoid that mistake in future.
I tarted this column with one of my favourite quotes by John Donne from his "No man is an island" piece. Donne was extremely well-travelled for his time, and maritime analogies often pop up in his writings. At about 1600, he wrote the following:
“Whoever loves, if he do not propose
The right true end of love, he’s one that goes
To sea for nothing but to make him sick.”
So I urge you, gentle reader, to go to sea not to be sick (which is common on tugs), or to lord it over your shipmates by not being sick, but to keep your eyes open, speak up when something is not right and potentially save lives in the process.
Unless we can start doing a better job of preventing serious accidents, it will be difficult to argue that humans are better than AI.
Sadly, there is a real danger that we might lose people with the ability to spot a potential hazard and take steps to prevent it. I refer, of course, to the advent of artificial intelligence (AI).
At a recent Splash conference in Singapore we are told that panellists, “wrestled with trust, accountability, skills and the risk of automating away the very experience the industry will need in future.”
It was my old friend Pradeep Chawla, a man who has dedicated his life to maritime safety and training, who asked the biggest unresolved question: “If AI removes the repetitive junior tasks through which generations of shipping professionals traditionally learned their craft, who becomes the expert capable of challenging AI in 20 or 30 years?”
One panellist said it was the best question he had received on any AI panel, but admitted he was unable to answer it. And that is the problem in a nutshell – we will introduce the technology simply because we can, and it will save money that is currently being spent on junior people, but it will be left to future generations to sort out the mess.
Splash summed it up adequately:
“That may ultimately be shipping’s most important AI problem: not whether machines can learn maritime expertise, but how humans will continue to acquire it.”
And unless we can start doing a better job of preventing serious accidents, it will be difficult to argue that humans are better than AI. Proponents of AI will ignore the thousands of tug jobs that are accomplished every day without incident, and point to the tiny percentage where humans screw things up.
We need to reduce the number of accidents, or the people who come after us may be condemned to going to sea merely to be sick.