Maritime organisations must adopt a cautious approach to integrating AI into their business operations and factor in high future costs and potential stranded assets, writes  Namrata Nadkarni, CEO & Founder, Intent Communications. Photo credit: pexels (pexels-pixabay-164652)

Artificial Intelligence (AI) is being aggressively promoted within various sectors around the world and maritime is no exception. Vessel design, port optimisation, condition monitoring, data analysis and research and more are all seen as ideal cases for AI integration. The ability to automate operations, accelerate processes by more than 10 times the speed it would take a human to perform a task, and of course, reduce running costs has a significant appeal for maritime leaders trying to make their businesses more profitable in turbulent markets. 

Additionally, agentic AI would be able to offer round-the-clock support without the need for breaks, which is a customer service dream if managed properly – and if the AI can be trained enough not to expose the company to a significant liability (warning, we are nowhere near that level of training as yet).

But, before we embrace AI as the all encompassing solution it is being positioned as, we must critically assess the business case for its use. I say this not as a technophobe, but as someone who wants our sector to thrive. While I think that AI has some wonderful – and specific – use cases, I also believe that a critical assessment is necessary, as widespread adoption could really impact businesses negatively in the long run. 

Cause for concern

AI, in its current deployment phase, is a massive loss leader, with costs artificially repressed by tech giants and venture capital firms that are pouring money into capturing market share and establishing platform dominance to create a sector-neutral reliance on these technologies. When the bill comes due, the costs will be passed on to the consumer. In some cases, it will be worth paying for access as the technology is truly improving operations. But in many of the current implementations, the main driver for use of AI is to lower operating costs – which will be reversed when the subsidies end.

The assumption of higher costs is not baseless. Large Language Models and generative AI are high level energy consumers and energy prices are already on the rise (particularly given the war with Iran, which is an oil and gas producer). As these costs increase, they will need to be passed on to users in order for the AI supplier to stay viable. Furthermore, we are likely to see emission related taxation, as well as other regulations which will inevitably add layers of compliance costs.

It is also worth pointing out that the infrastructure – specifically the high-performance microchips needed to create the data centres – are already seeing tightening across the supply chain, with the next few production cycles already purchased. This will inevitably create a spike in prices, particularly as consumers will also be competing for this hardware. 

Finally, and perhaps most importantly, the shareholders of these AI behemoths will eventually demand a return on their massive investments. When that pivot from ‘growth at all costs’ to ‘profit extraction’ happens, the pricing models for the AI services underpinning maritime and other administrative operations will change drastically. The monthly/annual fee model will almost certainly pivot to a consumption based model, with high users needing to cost this into their annual budgets.

The risk for maritime businesses is not the technology itself, but how we are integrating it. If we build our operational resilience around AI at its current subsidised price point, we are building on sand.

Assessing the landscape

Maritime companies are understandably eager to reap the benefits of this new technology and we are already seeing swathes of job losses as manual processes are replaced with AI workflows, assuming that the current cost efficiencies are permanent. We should not automatically assume AI is a cost-saver. Instead, we must stress-test our adoption strategies. 

Maritime leaders must ask themselves: Is this integration profitable even if the cost of the technology doubles? Are we willing to absorb these additional costs? If the answer is no, then stripping away the human redundancy that acts as a safety net is a strategic error.

There is also a growing risk of stranded assets where business-critical functions centered on a specific AI supplier run the risk of becoming cut off from after-sales support and service if the company goes out of business or decides that a specific product is not profitable enough. 

Preparing for the inevitable

I believe that the best way forward is to adopt a considered approach to AI, treating it as a specialist tool rather than a general-purpose crutch. As part of Intent Communications’ ongoing campaign to demystify digitalisation and AI in maritime, my advice to maritime leaders is to set strict parameters. 

We must identify the critical functions where AI delivers such undeniable value that it is worth paying a premium for. In these high-value use cases, the ROI will likely survive a price hike. For everything else—widespread adoption for administrative ease or marginal efficiency gains—we must be far more cautious. We need a clear plan of action for what happens if costs spiral beyond our expectations. We need to know exactly how much we are willing to pay, and at what price point we pull the plug.

We cannot afford to be vulnerable to costs beyond those anticipated. Let us use AI where it makes us stronger, but let us not hand over the keys to our operational stability without reading the fine print. The ‘free trial’ period of the AI era is ending; make sure your business can afford the full subscription.