“The winners will not be the companies with the most AI.”

AI is increasingly handling repetitive work such as data analysis, reporting and monitoring because it is exceptionally good at processing large volumes of information, identifying patterns and automating routine tasks. At BSM, AI is becoming more and more embedded in our processes and workflows, ranging from drydock planning solutions that support maintenance decisions to chatbots that help identify fuel-efficiency improvements and tools that automate elements of vessel inspections. AI really helps connect information across functions such as procurement, crewing, compliance and finance, providing predictive insights and reducing manual data handling. This is good. It truly makes our processes more efficient.

But shipping is full of exceptions because nature and human behaviour are full of exceptions. The world is never entirely predictable, and into this imperfect world comes AI, using data and pattern recognition to identify opportunities for optimisation. Yet weather changes, port conditions, equipment behaviour, regulatory interpretation and commercial pressures do not always fit neatly into a model, nor do we – the so-called human element.

That is why AI systems can identify risks and provide recommendations. They may even come across like a sparring partner. But they will always only be as good as the data they will receive and people must remain responsible for making the final call.

As AI systems become more capable and operate with greater autonomy, they will increasingly give the impression that they have everything under control. And so, the most important skill will be maintaining a mindset of critical thinking, grounded in expertise and experience. We must not allow ourselves to be seduced by the confidence with which AI presents its answers.

People need to understand data and system behaviour. They must be able to assess whether a recommendation makes sense in a specific operational context. In particular, they need to excel at managing exceptions and understanding the bigger picture: how operational, technical, regulatory and digital considerations interact.

There will also be a growing need for AI interaction skills: knowing how to query systems effectively and recognise their limitations. As vessels become more connected, cybersecurity awareness will become even more important.

The good news is that these are not entirely new skills for maritime professionals. Seafarers, in particular, operate in an environment that constantly requires sound judgement and situational awareness. The challenge is therefore less about learning a new way of thinking but more about maintaining the same level of vigilance when working alongside AI, especially when its recommendations appear entirely reasonable on screen.

The biggest concern is the mismatch between the speed of technology adoption and the speed at which training and competency frameworks evolve.

Traditionally, maritime education and regulatory frameworks develop over years rather than months. AI capabilities can change significantly within weeks.

Many around the globe are becoming aware of this urgency. Recently, a coalition of 29 European states and the International Association of Classification Societies called on the IMO to begin developing dedicated standards for AI systems used in shipboard navigation and communications equipment.

From an industry perspective, that is encouraging. Shipping is not only full of exceptions but also full of fragmentation. Just look at decarbonisation. Companies are dealing with EU ETS, UK ETS and FuelEU Maritime while everyone is still waiting for the IMO's Net-Zero Framework to become a reality. With AI, we might have the opportunity to set global standards much earlier and avoid a similar patchwork of requirements.

The pace of technological change means upskilling cannot be classroom-first. Learning to work with AI requires speed and continuous development. Most of the learning will need to happen on the job, as part of the daily work. It should focus on data literacy and understanding when to trust and when to challenge AI-generated recommendations.

Companies will need to accept the continuous evolution of their operating frameworks, defining what AI can do and the role people must play. We did this at BSM and have established an AI policy and governance process to ensure AI solutions are assessed, approved and used responsibly, with risk controls and human oversight.

The winners will not be the companies with the most AI. They will be the companies that best combine human judgement with AI at scale.

Leaders should ensure that humans and AI work together within defined boundaries. They must lead by example and embrace and promote continuous learning. As AI becomes more capable, it must always be clear that accountability remains with people.