What the IMO MASS Code Assumes About Autonomous Ships
The IMO's 111th Maritime Safety Committee (MSC 111) adopted the international safety code for Maritime Autonomous Surface Ships (MASS) — the MASS Code, Resolution MSC.595(111) — on 21 May 2026. The Code doesn't replace SOLAS; it supplements it as a non-mandatory, goal-based instrument: each chapter states a goal and functional requirements, not a prescribed technical solution. It applies to SOLAS Chapter I cargo ships and their associated Remote Operations Centres (ROCs), and only where a flag Administration decides existing rules are impractical or insufficient for a given autonomous function.
The most striking assumption is that the master's responsibility doesn't disappear. The Code accepts that a human master need not be aboard, but requires that the master be able to intervene at any time. Every autonomous function must go through its own risk assessment and be submitted to, and approved by, the flag Administration individually — there is no single line-item approval that says "this ship is autonomous." The sum of function-level approvals is the ship's actual autonomy envelope.
An Ontology Is Not a Blueprint — It's a Declaration of Accountability Paths
Because a goal-based code doesn't prescribe technical solutions, the operator has to declare, for itself, what each autonomous function connects to and who holds the authority to intervene. Scatter that declaration across separate documents and consistency breaks the moment a new function is added. Model it as an ontology instead — function, risk-assessment version, ROC binding, intervention authority, and flag-state approval status as one connected graph — and the question "can we turn this function on right now" gets answered instantly, without hunting through paperwork. Borrowing the entity-relationship skeleton of adjacent standards such as ASAM OpenX or IEC 63278 (Asset Administration Shell) means you don't have to design a schema from scratch.
From Risk Assessment to ROC Operations: Building and Running a MASS Ontology
Planning-stage targets should be numeric. Set the lead time from risk-assessment submission to flag-state approval for a single new autonomous function at 60 days or less, set the initial ship–ROC compatibility survey pass rate at 100%, and flag any function that misses either target into an automatic hold state inside the ontology.
The most common failure is scope creep: a ROC operator remotely controlling a wider set of functions than the risk-assessed, approved scope allows. Without a declared per-function approval boundary in the ontology, this overreach stays invisible until an incident investigation finds it.
The second failure pattern is link loss. If ROC–ship communication drops, or the cybersecurity boundary required under Annex section 9.8 is breached, the system needs an immediate branch to local control. The operating bar is to confirm, within 5 minutes of detecting a communication loss, whether onboard intervention is possible — and if it can't be confirmed, to fall through to a safe-stop procedure.
Before any deployment, run recurring ship–ROC compatibility re-check scenarios, log per-function approval state, risk-assessment version, and intervention-authority holder as standard fields, and mask location and personal data in communication logs so the flag-state submission copy and the internal operations copy stay separate.
The Experience-Building Phase framework IMO is set to establish by December 2026 is the channel for feeding real-world anomalies back into future code revisions. If the ontology already accumulates incidents and anomalies alongside each function's risk tier, there's no new dataset to build when that framework calls for submissions.
A mandatory code isn't expected to take effect before 2032 at the earliest, which makes today's voluntary window a grace period for starting with a lightweight schema and refining it as functions are added. Defer that work and the full cost of data migration lands all at once when the mandatory phase begins.
The Checklist to Use Right Now
Instead of prescribing technology, the MASS Code asks operators to prove, function by function, that a risk assessment exists and that a human can intervene. Tie function, approval, and intervention authority into a single ontology graph, and you can respond in minutes to the two baseline failure modes — scope creep and link loss — while spreading out the cost of the eventual mandatory transition starting now, well ahead of 2032.
References
IMO adopts first global Code for autonomous ships — International Maritime Organization
Ask AI about this article
The assistant has read this article. Ask anything — it answers from the text and says so when something isn't in it.
Loading the chat…