FEATURE
AI Sets Sail
Artificial intelligence is already reshaping the maritime industry, and real-time systems that can monitor and adapt instantly at sea are on the horizon.
Written by Lina Zeldovich Illustrations by Andreea Dumuta
Photo: Getty
IT’S 2030, AND A CARGO SHIP is sailing fast on the high seas. While the captain is standing on the bridge looking at a computer screen, an AI assistant interrupts: A pod of orcas lies ahead at one o’clock. Should the ship slow down or reroute?
Within seconds, the system evaluates the currents, calculates the timing, and estimates fuel use before charting a new course for the captain.
This scenario may sound like a scene from Star Trek, but it could soon be a reality, according to experts who work in the shipping and offshore fields. Just like everywhere else, AI is quickly becoming a gamechanger in the marine industry, revolutionizing the way ships are designed, built, maintained, and operated.
And, once the ships are outfitted with more advanced sensors, AI could be able to detect marine life, analyze currents, winds, and other factors—and help chart optimal routes under the constantly changing conditions of the high seas. Such developments would not only impact the length of the journey, but also improve fuel use and make shipping more cost-effective, sustainable, and environmentally friendly.
“In the marine industry, there are some huge wins that will have a big impact on the way that goods are shipped around, in the way that people are carried around, and also on the greenhouse gas emissions,” said Adrien Desjardins, a professor and Seaspan chair in robotics for marine vessels at the University of British Columbia. Desjardins moderated an AI and marine sustainability panel at a recent ASME conference. “I think the big picture is that there will be a transformational change,” he added. And that change will affect just about every aspect of the marine industry.
Turbocharged Digital Twins
Building ships is time-consuming and expensive, and that’s where AI is helping already. In ship hydrodynamics, water resistance is generated by both the bow and the aft, but the types of resistance and the factors influencing them differ at each end. The size and shape of the hull play a crucial role in how a ship moves through the water and how much fuel it uses, said Rajeev Jaiman, associate professor in the University of British Columbia Department of Mechanical Engineering. And because ships aren’t mass-produced like cars or planes, and are usually built to the customers’ specific parameters, if mistakes are realized after the ship is out of the yard, they are very costly.
To tackle the problem, engineers are deploying AI to design more efficient hull structures, automating design processes that were once manual and experiential. Analyzing complex parameters like hydrodynamic resistance—the resistance of water caused by turbulence or changing flows as ships move through it—has been a laborious process to test because it requires manual computations. AI models can crunch through the numbers quickly, allowing engineers to identify optimal hull forms, reduce drag, and boost overall sailing performance, all while optimizing sailing for fuel efficiency, cost, and greenhouse gas emissions reduction. “AI can definitely provide much better decision-making power for both design and operations,” Jaiman said.
And AI can also help engineers create better digital twins—virtual replicas of physical vessels, used for analysis, simulation, and prediction. Digital twins aren’t a new concept, and shipbuilders have been using them to analyze future ships’ performance in design stages. “Where the AI comes in is potentially the ability to do that at a much finer level of granularity, and more efficiently computationally,” Desjardins said.
“You can think about using AI to immediately scan through a 2D drawing and identify that all the fire extinguishers are there. Or, if I take a video during inspections of a cargo tank on a ship, AI can automatically identify cracks or corrosion.”
—Eric Vanderhorn, manager of product development at the American Bureau of Shipping
Photo: Getty
AI models can be trained on older vessels to pinpoint the most optimal combinations. By consolidating current and historical data, these AI-enabled digital mimics can simulate real-world conditions, allowing engineers to monitor, diagnose, and optimize performance before the ship is built without ever setting foot onto a shipyard. With AI-driven digital twins and the ability to test design variations in a fraction of the time, engineers can identify the most efficient and environmentally friendly configurations before a ship ever hits the water. “AI can quickly come up with new hull forms,” Jaiman said. “Different curves, different parts of the ship can be quickly redesigned using the generative AI-based toolbox.”
Jaiman foresees that these digital twins will move out of design stages into the real world. “We used to call them fixed digital twins,” he said, but now he’s envisioning dynamic digital twins, operated by generative AI live at sea, which every ship will eventually sail with. “They will be talking to each other real time, sending information back and forth dynamically,” he said.
AI may also help with predictive maintenance by reading data from sensors installed on critical machinery while the ship is sailing, proactively alerting crews if emergency maintenance is needed.
That will be particularly important for ships sailing in harsher, more unpredictable environments as more Arctic routes will become navigable. “As the Northern Sea route is opening up, a lot more ships will be going across the Arctic environment,” Jaiman said. “AI and the digital twins technologies will be very, very instrumental for Arctic shipping.” That real-time feedback will enable the ship to adjust to the changing conditions quickly, whether altering the course to avoid disturbing wildlife or adapting to changing currents, wind, and ice, while using the fuel more efficiently.
Better Safety and Maintenance
Eric VanDerHorn, manager of product development at the American Bureau of Shipping (ABS), which sets and enforces safety standards for the design, construction, and maintenance of marine and offshore structures, said the organization has already adopted AI into some of its daily operations. With 200 offices in 70 countries around the planet, ABS’s network of surveyors and engineers oversees between 12,000 and 13,000 floating assets worldwide, including ships, barges, and offshore platforms—a feat where AI certainly comes in handy.
Up until recently, checking a vessel’s machinery required extensive human involvement. The inspectors would come in and check the engines, generators, and other moving parts, take a lube oil sample to be analyzed in the lab, ultimately producing a report about the ship’s health status that would either recommend certain maintenance or give a green light to sail for another 10,000 or so running hours.
Today, this is changing. “What we’re seeing now is the emergence of AI-driven approaches that are taking sensor-based measurements coming from these systems,” VanDerHorn said.
Photo: Getty
“As the Northern Sea route is opening up, a lot more ships will be going across the Arctic environment. AI and the digital twins technologies will be very, very instrumental for Arctic shipping.”
—Rajeev Jaiman, associate professor in the Department of Mechanical Engineering at the University of British Columbia
The data is loaded into computer systems, whether onboard or in the cloud, where all the calculations and analysis are happening, and then the AI presents its suggestions. “Instead of having to go through that human [subject matter expert], AI is coming up with, ‘here’s my recommendation,’ and then a human needs to go and make a decision based on that recommendation,” he said.
There will be more work for AI in design and maintenance, VanDerHorn speculated. “We have engineers in the office who are reviewing the drawings, and we have surveyors in the field who are actually going out to these assets, both in the construction yards as well as in the ports, to verify that it’s being built to the approved drawings, and then that it’s being maintained over time,” he said. “So you can think about using AI to immediately scan through a 2D drawing and identify that all the fire extinguishers are there. Or you can think about, if I take a video during inspections of a cargo tank on a ship, AI can automatically identify cracks or corrosion.”
Today’s AI systems mostly focus on detecting various anomalies, but the future ones will be better at diagnostics—zeroing in on the underlying reasons. Another step forward will be the “prognostic” abilities, as VanDerHorn describes them. “This is what’s happening, and this is how long I think you have,” he said.
Noise Reduction
Another big challenge in maritime shipping is the noise that ships make as they move through the water, which affects wildlife. The constant, low-frequency whirr that comes from the vessels—which grew significantly as ships increased in numbers—interferes with marine animals’ echolocation, disrupting their communication, navigation, and foraging. Some studies suggest that it can cause hearing loss and lead to behavioral changes. In certain protected areas, ships must reduce speed to make less noise, but that doesn’t apply to high seas.
The ship’s propellers are the primary culprits of the disturbance, said Jaiman. That noise is a result of the phenomenon called propeller cavitation, in which the rapid formation and collapse of steam-filled bubbles from the rotating propeller blades generate too much acoustic energy, which travels through the water as a soundwave. Other noise sources from ships include the engine and hydrodynamic flow, but cavitation is the most significant contributor to the acoustic disturbance ships cause to marine life. “Cavitation makes a lot of bubbles, and when they collapse, they make a loud noise,” Jaiman explained. “That noise can go inside the water up to 10 to 100 kilometers, and it’s very loud.”
Photo: Getty
Photo: Getty
AI can help make ships quieter by testing different propeller shapes and operating speeds to find designs that create fewer bubbles—another strong case for digital twins, given the speed at which the technology can crunch through numerous variations. “We can take physics models, build machine learning models on top of that, and those machine learning models will go inside the AI,” said Jaiman, who has been working on creating this functionality for the past seven years. “So we have already built this kind of an AI digital twin toolbox.”
The next step would be to use AI-powered real-time digital twins to spot marine wildlife and reduce the noise. “If we can, in real time, predict the noise generated by a propeller at a location where you think whales or other mammals might be, then you can adjust your speed,” Jaiman said.
The functionality would be invaluable at large ports and beyond, and it could be a significant step forward for ocean health.
Navigating Constraints
No such dynamic, real-time digital twins exist today, and it will take time and money to build them, experts acknowledge. The bottleneck is computing power and money. “To build a physics model for the whole ship, we need a lot of computing resources, and in engineering, we don't have that,” Jaiman said. “The marine industry is very economical and low cost. We don’t have billions of dollars to spend on hardware... With limited resources and computing hardware, it’s taking time.”
Another challenge is availability of data. AI models benefit from having as much info as possible. But not every ship builder or operator wants to share their data with competitors. And they may have other reasons to keep information private—such as when some vessels aren’t performing at their best.
“That is a very legitimate challenge, availability of information,” VanDerHorn said. But he is hopeful that once people start seeing the benefits—improved safety, reduced costs, and environmental wins—they will become more open to sharing their data.
Despite the challenges, the consensus is that the AI ship has set sail. Reaching its full potential will depend on advances in computing power and increased investment, but the course is clear: toward more connected systems at sea.
Lina Zeldovich is a science and technology writer based in Woodside, N.Y. Her most recent book, The Living Medicine: How a Lifesaving Cure Was Nearly Lost—and Why It Will Rescue Us When Antibiotics Fail, was published in October 2024.
Photo: Getty
© 2026 The American Society of Mechanical Engineers. All rights reserved.