The maritime sector is experiencing a transformative shift as Artificial Intelligence redefines how ships are operated, maintained, and navigated. Maritime is embracing AI with open arms, driven by the need for increased efficiency, safety, and sustainability. AI’s ability to process vast amounts of data and make real-time decisions is helping optimise voyages, reduce fuel consumption, improve navigational safety, and ensure better reliability across the board. Let’s take a closer look at AI in the maritime industry.
The maritime Artificial Intelligence market is expanding rapidly. Valued at £4.13 billion in 2024, it is projected to grow at a 23% Compound Annual Growth Rate (CAGR) over the next five years, according to a recent report from Lloyd’s Register. In the last 12 months alone, 420 organisations have adopted AI technologies in maritime, a significant rise from 276 in 2023. Start-ups and SMEs are leading the charge, making up 63% of the AI tech suppliers in the sector. A prime example is Orca AI, a (as they call it themselves) ‘fully automated lookout on the bridge’. Orca AI secured £23 million in May 2024 to enhance its platform, improving voyage safety and reducing CO2 emissions by 170,000 tonnes annually. This growth shows just how quickly Artificial Intelligence is becoming integrated into the maritime sector. But how exactly is it transforming the industry?
Unlike traditional methods, AI systems continuously learn and adapt. How? The answer is machine learning. Imagine a machine that doesn’t just follow instructions but learns from experience becoming smarter over time. That’s the magic of machine learning, a branch of Artificial Intelligence that’s revolutionising industries and everyday life. Think of it as teaching a computer to recognise patterns, make predictions, or even develop solutions based on past experiences. Unlike traditional software, which operates based on explicitly programmed instructions, machine learning algorithms get smarter the more data they process, gradually improving their accuracy, efficiency and self programming. In simple terms, machine learning algorithms are sets of instructions that help computers learn from past data. The more training data the system receives, the better it can predict future outcomes. This “learning” ability (the quotes are not to be ignored, as what we call ‘learning’ is mostly statistical induction) is what enables Artificial Intelligence to handle complex tasks — from diagnosing diseases to predicting the stock market.
IBM, in collaboration with NASA, has released a new open-source AI foundation model called Prithvi WxC, designed for weather and climate-related applications, where ‘open-source’ simply means the software is freely available to everyone, unlike proprietary software that requires a license. A ‘foundation model’ refers to an Artificial intelligence system that’s initially trained on a broad task, and can then be fine-tuned or customised to perform a variety of specific tasks — think of it as a ‘base’ model that learns general patterns, and then can be ‘adapted’ to handle more specialised tasks, much like how a general knowledge meteorologist might be trained to focus on a specific f ield like marine weather. Trained on 40 years of historical weather data from NASA’s MERRA-2 dataset, the model is built on advanced AI architectures, including a masked autoencoder, a type of AI model designed to learn from incomplete data. A masked autoencoder works by ‘masking’ or hiding part of the information and then learning how to predict or fill in the missing pieces — think of it as a puzzle-solving Artificial Intelligence that uses the pieces it has to guess what’s missing. These sophisticated techniques allow it to handle complex spatial-temporal data, making it one of the most advanced tools in maritime forecasting. It also has a unique training process that teaches it to predict missing weather data, mimicking the forecasting process. Nonetheless, it is also designed to run efficiently on desktop computers, offering accessibility even without massive supercomputing resources.
While Artificial Intelligence’s impact on weather forecasting helps optimise voyage planning, its applications extend to real-time navigation, where it plays a crucial role in collision avoidance, another area where machine learning-based Artificial Intelligence is making significant strides. Advanced AI-powered navigation systems integrate data from radar, GPS, and Automatic Identification Systems (AIS) to detect nearby vessels, predict their movement, and autonomously adjust a vessel’s course to avoid collisions. These systems help ships make real-time decisions, even in low visibility or crowded waters. One example is SEA.AI. SEA.AI uses the latest camera technology in combination with artificial intelligence to alert crews early and reliably about objects on the surface of the water — objects that might otherwise escape conventional systems like Radar or AIS. Whether it’s unsignalled craft, floating obstacles, buoys, inflatables, kayaks, or even persons overboard, SEA.AI can detect and classify these potential hazards in real-time. The ability to reliably detect and respond to obstacles on the water, particularly those that are not easily picked up by traditional sensors, is crucial in preventing accidents and improving navigational security.
As Artificial Intelligence and other technologies advance, the role of the skipper is also evolving. Navigational tasks that once required constant human attention — like monitoring foils or adjusting sail settings — are now assisted or even automated by Artificial Intelligence, freeing up sailors to focus on more strategic decisions. The evolution of autopilot systems in high-performance sailing boats exemplifies this trend. MADBrain, a cutting-edge autopilot system developed by Madintec, is a champion of this. Powered by artificial intelligence, MADBrain continuously optimises a boat’s performance, autonomously adjusting the rudder based on real-time data, such as speed, position, wind conditions, and wave patterns. As it learns from each sailing experience, the system improves its reactions and performance, ensuring the vessel stays on course even in challenging conditions.
The integration of systems like MADBrain and SEA. AI is revolutionising how modern boats are sailed. For this reason, even the insurance industry is embracing AI. Insurers are increasingly turning to AI to assess risks, streamline underwriting, and optimise claims management for both large vessels and smaller crafts. AI is helping insurers predict potential risks more accurately, set more appropriate premiums, and even anticipate claims before they occur. This data-driven approach is revolutionising risk management in both commercial shipping and recreational boating. Some insurers are even using AI to monitor ships in real-time, offering proactive maintenance suggestions to prevent costly repairs and operational disruptions.
Interestingly, many innovations that are transforming how we think about commercial maritime operations get tested in the world of sailboat racing — where the goal is not just to win but to push the boundaries of technology. Teams invest millions in research and development, and every race introduces new advancements in boat design, performance, and safety. The AI tools developed for these high-speed, highrisk races are increasingly being integrated into commercial vessels, autonomous shipping, and even marine surveying. For instance, the same AI that helps race teams optimise their strategies in real time is being adapted to improve navigation systems on larger vessels. What starts in the high-stakes world of racing often filters down to influence commercial shipping, recreational boating, and other maritime sectors, showing the intersectionality of Artificial Intelligence.
The latest editions of the America’s Cup, for example, have represented a proving ground for AI’s potential in data analysis and decision-making. During New Zealand’s victorious 2021 campaign, AI was crucial in optimising every aspect of the team’s strategy. As one team member put it, “We had the best sailors, the best boat, and the best AI.” This collaboration between top-tier Artificial Intelligence and expert sailors allowed the team to process vast amounts of real-time data — from wind patterns to competitor movements — and make split-second adjustments to their tactics. The result? Faster boats, smarter decisions, and ultimately, victory. As mentioned, the AI tools developed for the America’s Cup have farreaching applications, with insights and innovations f lowing into larger commercial vessels, autonomous ships, and even port management systems.
And so it goes with the Vendée Globe, considered one of the toughest solo sailing races on the planet. Yachts like the IMOCA 60s are equipped with sophisticated data systems powered by AI, which autonomously adjust key operations such as rudder angles. This allows sailors to save crucial mental and physical energy rather than micromanaging every aspect of the boat’s performance. For example, the aforementioned MADBrain autopilot system, which continuously adjusts the course based on changing conditions like wind, waves, and speed, and SEA.AI, are AIdriven technologies that are laying the foundation for the future of all vessels. These same technologies, tested and perfected in racing, are now set to become standard practice.
Something surveyors might want to keep an eye on is the intersection of AI and robotics. AI-powered robotics could revolutionise boat inspections, a crucial aspect of maintaining the integrity and safety of vessels. Traditionally inspecting ship hulls and other critical parts of a boat required humans to work in challenging and sometimes hazardous environments. This process was not only time-consuming and expensive but also posed significant risks to personnel. Now AI-powered remotely operated vehicles (ROVs), whether aerial or subaqueous, are transforming the way inspections are carried out. AI-enhanced drones are autonomous robots equipped with a range of advanced sensors and high-definition cameras designed to inspect or from top to bottom. The combination of AI and robotics allows these vehicles to not only gather visual data but also to process it on the fly, identifying issues that might otherwise go unnoticed. Using sophisticated AI algorithms, these vehicles can detect cracks, corrosion, structural weaknesses, or damage to the hull, as well as monitor the condition of vital systems like rudders and propellers. The AI can flag any abnormalities or patterns that suggest emerging issues, prompting early maintenance before they escalate into major problems.
Yet, despite all the promises of Artificial Intelligence, its use in the maritime sector still face several challenges, Data quality is one of the biggest hurdles. AI systems rely on accurate, timely, and detailed data to function properly. Without high-quality data, AI struggles to make decisions. To make autonomous vessels safer and more efficient, experts like Dr. Hideyuki Ando emphasise that clean data is non-negotiable. For companies like NYK Line, investing in better sensors and infrastructure to collect robust data isn’t just a short-term fix — it’s the foundation for developing advanced autonomous systems. But there’s a catch: the maritime industry lacks standardised data compared to other sectors, making it difficult to train AI models effectively.
AI also faces a trust barrier. Crew members are naturally wary about letting AI make critical decisions without understanding how it works. Stena Line’s Michael Ljunge reveals that the company had to reframe their AI system — Captain’s AI — to be seen as an intelligent assistant rather than a replacement for human judgment. This shift in perception has helped increase acceptance onboard, as the AI now supports the captain’s decisions rather than trying to replace them.
While the potential benefits of AI in the maritime industry are clear, it’s important to recognise that these technologies are still in their early stages of deployment. AI systems, especially in complex and unpredictable maritime environments, have limitations. Challenges such as data quality, system reliability, and the need for human oversight in decision-making are ongoing considerations. As AI continues to mature, industry leaders will need to navigate these hurdles to fully realise its potential, ensuring a harmonious balance between automation and human expertise.
As AI continues to evolve, its impact on the sector will likely only grow, as the maritime industry will want to stay ahead of the curve in an increasingly tech-driven world. As we have seen, the lessons learned and innovations tested in the competitive world of race boats are directly influencing the future of all maritime operations. Although or from top to bottom. The combination of AI and robotics allows these vehicles to not only gather visual data but also to process it on the fly, identifying issues that might otherwise go unnoticed. Using sophisticated AI algorithms, these vehicles can detect cracks, corrosion, structural weaknesses, or damage to the hull, as well as monitor the condition of vital systems like rudders and propellers. The AI can flag any abnormalities or patterns that suggest emerging issues, prompting early maintenance before they escalate into major problems.