Where Is the Captain? Eight Weeks Inside the Voluntary Phase

Week 37 | August 2026

"There should be a human master responsible for a MASS, regardless of the mode of operation, and the master should have the means to intervene when necessary." That's clause II/8.7.3 of the International Code of Safety for Maritime Autonomous Surface Ships (Code), adopted in May and in effect since 1 July of this year. Read on and it gets more interesting. The master "may be located physically on board or at an ROC." Command can transfer "to or within an ROC, or between ROCs." One master, one ship, at any moment, wherever that master happens to be sitting that day, in whichever flag state.

So: where is the captain, physically? Eight weeks into the voluntary phase, nobody's actually had to find out.

There's a specific reason for the silence, and it has a cost.

The Code took effect 1 July, but the Experience-Building Phase, the framework that would give "voluntary adoption" any operational meaning, doesn't actually get built until MSC 112 in December. Until then a flag state has nothing concrete to opt into. It can build a national regime. It cannot yet point to an agreed evidence standard and say it's running vessels under the Code.

The open registries have noticed. Bureau Veritas's MSC 111 wrap flagged that Panama, Liberia, the Marshall Islands and the Bahamas, the four flags carrying most of the world's tonnage, have all "expressed reservations." Singapore, arguably the most autonomy-forward port authority anywhere, recently pushed its flagship autonomous feeder tender back from 24 July to 21 August and softened the ask to "current or near-final versions." A remote operations centre is a required element of that project. Eight weeks in, the state best positioned to move first hasn't finished writing the proposal.

Chapter 2 is specific about when the Code bites. It applies to SOLAS cargo ships "including any associated Remote Operations Centre(s), which have systems and functions that enable autonomous or remote operations, when the Administration deems that compliance with other applicable instruments is impracticable or insufficient." The trigger is narrow and specific: it switches on wherever a ship's autonomous or remote functions make the rest of SOLAS impracticable to meet, crewed or not. 

The master provision holds regardless of mode. The ROC itself needs a MASS ROC Certificate, issued by the flag Administration, per Chapter 5.3. None of that is the gap.

Chapter 14.1.3 allows an Administration to treat a ROC as "a directly associated location to the navigational bridge or part of the machinery space… to ensure that the watchkeeping provisions of the 1978 STCW Convention and STCW Code may also apply to remote operators." Read that again. It's a narrow legal fiction, built for one purpose: letting shore-based watchkeepers count as qualified crew. It says nothing about enforcement or criminal jurisdiction. Nothing about what happens when the flag state wants to act against someone sitting in a ROC on foreign soil.

Its own preamble ties the whole Code to the United Nations Convention on the Law of the Sea. It never resolves what UNCLOS actually requires here.

The drafters looked directly at this question and closed it. An IMO working group briefing from the Code's development records "no need to further address concerns relating flag oversight over ROCs," leaning on the Legal Committee's 2024 finding that the ISM Code offered a workable model with no specific legal roadblocks. File closed, as far as the committee that wrote the Code was concerned.

Fifteen months later, the Norwegian Maritime Authority, a flag state actively licensing ROC operators rather than merely talking about it, commissioned an independent legal opinion on exactly this question. Published in February, it reaches the opposite conclusion. Article 94(1) of UNCLOS obliges a flag state to "effectively exercise its jurisdiction and control" over its own ships. Let the ROC sit in someone else's territory and the flag state may be unable to do that, which "could under certain circumstances amount to a breach." The ROC state, meanwhile, holds full territorial jurisdiction over anyone physically inside it, absent a treaty saying otherwise. No such treaty exists. The report's own prudent conclusion: put the ROC inside the flag state.

One body has called it solved. The flag state actually building the licensing infrastructure says otherwise.

The technology question is done. The drafting, for what it's worth, is detailed and genuinely goal-based. Nobody has settled who answers for the ship when something goes wrong and that ship’s ROC sits in one country, its flag is in another, and the vessel itself is transiting a third state's waters.

Ocean Infinity already runs remote operations centres across the UK, the US, Sweden and Australia. A fleet controlled from three jurisdictions is exactly the shape this problem is built for. Nobody's had to test the failure mode yet. December is when the next real signal arrives, not because anyone will have adopted anything by then, but because that's when the framework that makes adoption possible finally gets built.

Hyperion Systems printed a working hull in 40 hours in May. Nobody's heard from them since. Meanwhile the US Navy spent this northern summer proving something else: printed spare parts and drones can be mil-spec certified and built at sea without fuss. Next week we look at why one half of additive manufacturing in defence got boring and solved, and the other half, the actual hull, still can't find its way into a program of record. 

Since you have been, thanks for reading.

Cheers,

Mick

Ocean Tech Intelligence provides informational analysis only. Nothing in this publication constitutes financial, investment, legal, or strategic advice. Readers act on this content at their own risk. For full details see our Disclaimer.