Ships have been automated for half a century — autopilots since the 1920s, unmanned machinery spaces since the 1980s, integrated bridge systems for decades. So why does a Maritime Autonomous Surface Ship need its own regulatory code? Because the automation the rules were written for executes a human's decisions; the MASS makes them.
That distinction — execution versus decision — is the line the Code draws. The regulatory novelty isn't the technology, it's where the authority sits, and conventional automation never moved it.
Key findings
- Conventional automation executes within a human's decision frame — autopilot holds a course the officer chose. Established
- MASS autonomy makes decisions the human used to make — the authority, not the mechanism, is what's new. Established
- The regulatory gap is the decision-maker — existing rules assume a person decides; MASS needs rules for when software does. Established
- Degrees of autonomy define the continuum — from decision-support to fully autonomous, the code addresses the authority handoff. Established
- The liability and accountability chains differ — conventional automation's errors trace to its operator; MASS's trace to a system. Maritime-GPT analysis
The execution/decision line
An autopilot holds a course the officer set; an ECDIS alarm flags a danger the officer evaluates; the UMS alarms notify the engineer who's on call. In every case the human decides — the automation executes or alerts. A MASS's autonomy is different in kind: the software chooses the course, evaluates the encounter, makes the collision-avoidance decision. The regulatory problem the Code exists to solve is that 'who decides' is a legal category, not an engineering one — and conventional automation never crossed it. Established
The authority the rules assume
SOLAS, STCW, COLREGS, and ISM are all written around a human decision-maker — the master's authority, the officer's watch, the lookout's eye. Conventional automation fits inside that frame because it never holds the authority itself. A MASS breaks the frame: when the software decides, the regulation must either certify the software's judgement or redefine where the human's authority sits. That redefinition is the Code's actual content. Established
The accountability difference
A conventional automation failure traces to a person — the operator who mis-set it, the engineer who mis-maintained it, the master who mis-supervised it. A MASS's decision failure traces to a system — and the accountability question becomes architectural: was the system in its envelope? Was the override available? Who designed the autonomy's decision boundary? The liability chain that worked for execution doesn't work for decision — which is why a new code was needed rather than an amendment. Maritime-GPT analysis
What the Code adds
- <strong>Operational design domain.</strong> The bounded conditions the autonomy is authorized for — a concept conventional automation never needed. Established
- <strong>Fallback obligations.</strong> What the system must do when it exceeds its envelope — the degraded-safe behaviour. Established
- <strong>Remote-operations governance.</strong> The framework for human authority exercised from ashore. Established
- <strong>Assurance requirements.</strong> The evidence standard for a system that decides — the software's certification, not its hardware's. Established
What we know
- Conventional automation executes; MASS decides — the authority is the difference.
- Existing regulation assumes a human decision-maker; MASS breaks the frame.
- The Code's new concepts (ODD, fallback, ROC) address the decision authority.
What we don't
- Where the boundary between 'automation' and 'autonomy' legally sits in ambiguous cases.
- Whether existing vessel automation retrofits will need MASS compliance.
- How the Code handles hybrid modes (autonomous sometimes, conventional others).
Being tested
- Regulatory interpretation of the execution/decision boundary.
- MASS-degree classifications in practice.
- Assurance frameworks that span automation and autonomy.
Maritime-GPT analysis
The MASS-vs-automation question is why the Code exists at all: the industry has had automated ships for fifty years, but never one whose software held the decision authority. The Code isn't about new technology — it's about new accountability. Conventional automation is a tool the master uses; MASS autonomy is a decision-maker the master supervises. The regulation's novelty isn't in the ship — it's in who answers for it. Maritime-GPT analysis
This is a research feature, not operational guidance. Simulation results are reported as simulation; demonstrations are not deployments; company claims are attributed as such. Verify safety-critical decisions against official nautical publications and flag-state requirements.
Sources & evidence
The code whose novelty this article explains.