Autonomous Ships and the MASS Code

Strip away the headlines and an autonomous ship is a simple proposition with hard consequences: move some or all of the humans out of the control loop, and replace what they did with software, sensors, and a communications link. Everything interesting, technically, legally, and for this course, lives in that replacement.

Degrees, not a switch

Autonomy is not one thing. The IMO's work on Maritime Autonomous Surface Ships (MASS) frames it as four degrees. Degree 1: a crewed ship where automated systems support decision-making, seafarers ready to take control. Degree 2: a remotely controlled ship with seafarers on board. Degree 3: a remotely controlled ship without seafarers on board. Degree 4: a fully autonomous ship, whose systems decide and act themselves. That scale comes from IMO's 2018 regulatory scoping exercise and remains the industry's shorthand. The Code IMO adopted in 2026 (below) works differently: it applies to individual autonomous or remote functions and to the modes of operation a ship is designed for, even with crew on board for other duties, and it keeps a human master responsible for the ship whether or not that master is aboard.

Two things follow from the scale. First, autonomy is already here at the low end: track control, collision-avoidance aids and unattended machinery spaces put much of the world fleet near Degree 1. Second, the hard line is not between manual and automatic but between Degrees 2 and 3: the moment the last seafarer steps ashore, every function a human quietly guaranteed, lookout, firefighting, resetting a tripped breaker, must be engineered instead. A single voyage can also cross degrees: autonomous ocean transit, remote control in approaches, a crew for port manoeuvres.

The remote operation centre

For Degrees 2 and 3, the bridge moves ashore into a remote operation centre (ROC): a facility where operators monitor one or several vessels and take control when the operation demands it. An ROC is not a video-game cockpit; it is a regulated workplace with the problems of any bridge, plus new ones: situational awareness built from cameras and sensors rather than windows, qualification and workload rules for operators watching multiple ships, and above all the communications link. That link is now part of the ship's safety envelope. Its availability, integrity and authentication are exactly as safety-critical as steering gear: a spoofed or severed link is a lost ship in a way no conventional casualty category quite captures. This is where the course's two halves fuse: remote operation is only as safe as its cyber security, and the E26/E27 architecture of earlier lessons is the foundation autonomy stands on. Ships must also behave predictably when the link fails: minimum-risk behaviours such as holding position or proceeding on a safe pre-agreed track are designed in, because "call the bridge" is no longer an option.

The rulebook: voluntary first, mandatory later

Conventions written for crewed ships assume watchkeepers, lookouts, muster lists. Rather than amend everything at once, IMO chose a two-step path, and the first step has now been taken. A non-mandatory MASS Code was adopted by the Maritime Safety Committee on 22 May 2026 (resolution MSC.595(111)) and has applied to cargo ships since 1 July 2026; under it, early projects and administrations gather structured experience in an experience-building phase that MSC 112 (December 2026) is to set up. The second step is a mandatory code, which MSC intends to adopt by 1 July 2030 for entry into force on 1 January 2032, a date the committee has called the earliest and said may be revisited. The sequence is deliberate: freeze requirements too early and you regulate guesses; too late and every flag improvises. DNV's explainer on ensuring the safety of autonomous shipping, written while the voluntary code was still ahead, lays out this trajectory alongside the practical safety questions, autonomy levels, ROC design, and the link between autonomy and cyber resilience.

Class societies are not waiting. DNV's AROS notations (autonomous and remotely operated ships) give builders and owners a certifiable framework for autonomous and remote functions now, graded by how far the human has moved from the loop: the same gap-filling role class played for cyber before the IACS requirements landed. ClassNK's technical journal review of the status of MASS and class initiatives surveys the projects already sailing, short-sea and inland vessels, harbour craft, trial container tonnage, and the safety-assessment methods used to approve them.

For working operators, autonomy will arrive the way automation always has: function by function, not fleet by fleet. The skills this course has built, reading resilience requirements, questioning suppliers, drilling for the failure of a digital system, are precisely the ones autonomy demands more of. The last lesson widens the lens: the same wave of technology-first, rules-later change is reshaping cargo risk and regulation across the industry, and it rewards the same habits.