How to stay in command in the age of AI

AI will not make seamanship obsolete. It will make human judgement more valuable, as tomorrow’s crews learn to interrogate data, challenge automation and take control when increasingly intelligent systems get it wrong. The latest chapter from our new SplashTech magazine.

The crew of 2035 will not be short of technology. The more difficult question is whether they will be trained to challenge it.

Asked to identify the single most important skill future seafarers will need that maritime academies do not teach today, SplashTech respondents repeatedly return to variations of the same idea: digital judgement. Not coding. Not becoming a data scientist. The ability to understand what an intelligent system is telling you, recognise when its inputs or assumptions are weak, and know when experience must overrule the machine.

The urgency is already visible onboard. Dualog’s recent survey of around 300 seafarers and senior officers, half of them masters and with an average of more than 20 years at sea, finds surprisingly little resistance to artificial intelligence. Half have already experimented with AI. Ten face-to-face interviews with seafarers and shore personnel reinforce the picture.

“Seafarers are already using it – they want help solving problems,” Silje Moan, chief commercial officer at Dualog, tells SplashTech.

The obstacle is not curiosity. It is trust, source control and information overload. Connectivity has improved so dramatically that the frustration is shifting from getting information to finding the right information among manuals, procedures, messages and corporate systems. Dualog finds that improving findability makes crew 2.5 times more likely to be satisfied.

Moan also points to the administrative burden sitting on the bridge. Captains, she says, spend more than 50% of their time on admin work. AI could remove a large part of that burden, but only if it works from approved company sources and gives officers confidence that its answer can be relied upon.

Judgement beats fluency

Gert-Jan Panken, general manager and vice-president at Inmarsat Maritime, calls the required skill “digital judgement”. Seafarers will work with automated systems, AI-generated recommendations, remote support and growing volumes of operational data. The important capability will be knowing when to trust the technology, when to question it and how to combine digital insight with professional experience. Cyber awareness, he adds, will become part of normal operational competence rather than a specialist subject.

Joy Basu, chief executive of Smart Ship Hub, reaches the same conclusion through data literacy. Future officers need the confidence to read, question and act on data rather than either obeying AI or ignoring it. “By 2035 the vessel will not be short of data; it will be short of people fluent enough to turn that data into good decisions under real-world pressure,” he reckons.

The skill is having the judgment and competence to know when not to rely on it

Patrick Ryan, senior vice-president and chief technology officer at ABS, argues that as AI takes over more of the operational layer, seafarers will move increasingly towards oversight. That demands deeper, not weaker, domain knowledge. An engineer needs enough understanding to spot an implausible maintenance recommendation; a navigator needs the fundamentals to identify the edge case the algorithm misses.

Keeping the old skills alive

Garry Noonan, director of innovation at Ardmore Shipping, puts the requirement more starkly: the future seafarer must retain “the ability to operate when the system isn’t there”.

The danger of increasingly capable digital systems is not simply that they fail. It is that human capability can atrophy while they work perfectly. GNSS can be jammed or spoofed, ECDIS can freeze and optimisation software can produce a recommendation that simply looks wrong. In each case, somebody onboard must recognise the problem and know what comes next.

Noonan argues that academies should embrace technology without allowing traditional fundamentals to be crowded out. Navigation, engineering knowledge and situational awareness become more valuable when crews are supervising automation because they provide the independent reference against which technology is judged.

Jamie Jones, managing director of GTMaritime, gives the concept a name: “cyber-seamanship”. By 2035, he argues, a ship will be a fully connected and partly autonomous node, creating emergencies that may arrive through a data connection rather than as smoke, flooding or machinery noise. A manipulated navigation feed or compromised control system might leave every screen looking normal.

Traditional emergency training has turned responses to fire, flooding and abandonment into instinct. Jones argues digital casualties require the same treatment: officers must detect compromise, isolate affected systems and recover control under pressure.

Learning to supervise the machine

Other contributors widen the definition beyond cyber risk. Henning Davies, chief executive of LedgID, argues for “digital governance”: the ability to manage and interpret a growing landscape of tools, alerts, indicators and automated processes as seafarers move from repetitive execution towards system supervision.

Fabian Fussek, chief executive of Kaiko Systems, focuses on structured observation. Ships will still depend heavily on what humans see, hear and experience onboard. The future skill is capturing that observation in a form that both machines and shore teams can interpret, while retaining the judgement to challenge a digital system when experience says it is wrong.

P K Mishra, managing director of IRClass, similarly stresses automation bias. Future training needs to cover data literacy, cyber hygiene, alarm prioritisation and human-machine teamwork, he argues. The efficient officer is not the one who trusts technology most enthusiastically, but the one who adapts to it intelligently.

This is where the maritime academy curriculum faces a difficult balancing act. The answer is not to add a generic AI module and declare the problem solved. Digital judgement has to be trained inside navigation, engineering, cargo operations, maintenance and emergency response, because that is where it will actually be exercised.

More than an IT skill

The survey also produces answers that remind the industry not to define the future seafarer too narrowly.

Stuart Ostrow of ShipMoney chooses financial confidence. Seafarers are increasingly paid through digital and prepaid systems while supporting families across several currencies and jurisdictions. Knowing how to protect, manage and move money, he argues, will become an important part of working life at sea and remains barely addressed in conventional training.

That returns the argument to Dualog’s findings. Seafarers are not asking primarily for more technology. They are asking for technology that helps them find answers, reduces administration and supports decisions. Moan expects the captain of the near future to have a knowledgeable assistant using the vessel’s own documentation and data, capable of responding to a situation rather than merely searching documents.

The successful seafarer of 2035 therefore looks less like an IT specialist than a well-equipped mariner: traditional competence reinforced by data literacy, cyber awareness, critical thinking and the confidence to interrogate automation.

The defining relationship will not be human versus machine. It will be the quality of the partnership between them. AI can search faster, monitor more signals and identify patterns no person could process unaided. The officer remains responsible for understanding the sea, the ship and the consequences when the model gets it wrong.

Shipping academies have spent generations teaching people how to take command of a vessel. Their next challenge is teaching them how to remain in command when the vessel itself starts giving advice.

source : splash247