Why shipyards need AI helpers

Building a ship is a logistical nightmare. Thousands of pipes, cables, and pieces of equipment must fit into a precisely designed structure. But the reality is that components often arrive late or are installed in a different order than planned — resulting in discrepancies that block further assembly. Fixing a single such error can cost tens of thousands of dollars and delay vessel delivery by weeks.

This is exactly the problem that researchers from University of Michigan and the Massachusetts Institute of Technology (MIT), in collaboration with Japanese institutions, set out to solve. Their answer: autonomous robots equipped with artificial intelligence that continuously compare the actual state of the ship under construction with its digital twin — and detect deviations early.

$6.2 million and a test block: How the project comes together

The project is funded by Japan's Ministry of Land, Infrastructure, Transport and Tourism with $6.2 million (approximately CZK 140 million). Oversight of the research has been taken on by the Monohakobi Technology Institute, which falls under Japanese shipping giant NYK Line. The research will run until early 2027.

The robots will be equipped with LiDAR sensors (laser-based distance scanning) and cameras so they can 3D scan the ship during construction. The AI model will then automatically compare the scanned "as-built" state with the digital twin — a precise computer model of the intended design. If it finds a discrepancy, it flags it immediately.

A co-pilot, not a human replacement

The most interesting aspect of the entire project is the philosophy the researchers are taking. The AI robots are not designed as replacements for human workers, but as a "co-pilot." When the system detects a problem — such as blocked access routes, pipes that no longer fit, or missing clearance for equipment installation — it suggests alternative solutions and explains their pros and cons.

The system can also recognize where it has incomplete sensor data, and in such cases it requests human intervention. "The goal is not to replace an experienced shipbuilder, but to give them a tool that helps them make decisions faster and more accurately," the researchers explain.

Training on simulations and alongside seasoned masters

To ensure the AI understands real-world decision-making in a shipyard, the researchers won't train it solely on simulated scenarios. Plans include interviews with experienced craftsmen from both the US and Japan so that decades of practical know-how are baked into the system. The resulting model should thus propose solutions that make sense not only technically, but also in the context of real shipyard operations.

The technology will be tested on a so-called "Shipbuilding Test Block" — a physical model of ship sections that can be reconfigured to simulate various construction phases and assembly problems.

A broader trend: AI and robotics are conquering the shipbuilding industry

The MIT and University of Michigan project is not an isolated effort. Japan recently announced that artificial intelligence and robotics will be a priority in its shipbuilding industry revival strategy. In early 2026, Italian shipyard Fincantieri formed a partnership with startup Generative Bionics to deploy a humanoid welding robot that works alongside people directly in shipyards. Hong Kong Polytechnic University (PolyU) also launched a collaboration with Cheoy Lee Shipyards to research AI robotics and green materials.

Another example of automation in the maritime sector is Spanish company bound4blue, which received DNV Type Approval certification for its autonomous eSAIL system — a sail that uses the suction effect of wind to propel ships without human operation. The system autonomously optimizes thrust based on weather conditions and, according to the manufacturer, can reduce fuel consumption by up to 10%. The technology is already being deployed by shipping companies such as Odfjell, Eastern Pacific Shipping, and Louis Dreyfus Company.

What this means for Europe and the Czech Republic

The Czech Republic may not be a shipbuilding powerhouse, but that doesn't mean such technologies won't have an impact on local industry. The principle of a digital twin and AI-driven deviation monitoring from the plan is transferable to any kind of manufacturing — from automotive to mechanical engineering to construction. In the Czech Republic, digital twins are already being tested at places like Škoda Auto and in several Czech Technical University (ČVUT) projects.

For European shipyards — whether in Poland, Germany, or Croatia — the success of the MIT project would provide an accessible blueprint for reducing rework costs and shortening production cycles. With growing EU requirements for reducing maritime shipping emissions (the FuelEU Maritime initiative), the pressure for more efficient production will only intensify.

The project will run until early 2027, and the first results from testing on the Shipbuilding Test Block are expected during 2026. If the technology proves itself, it could appear in commercial shipyards within a few years — and the principles it's built on might soon show up in Czech factories as well.