The debate asks the wrong question. "Will AI replace seafarers" frames it as a single switch — crewed or uncrewed — when the real change happens function by function. A ship isn't one job; it's navigation, watchkeeping, machinery maintenance, cargo work, compliance, repair, and command. Each automates differently, on a different timeline, with a different effect on the person who does it. Maritime-GPT analysis
And part of the answer is already written in regulation: the MASS Code keeps a human master responsible at all times, and STCW's reach is defined by seafarers being aboard — meaning the legal framework assumes humans stay, relocated, not removed. Established
Key findings
- Watchkeeping is the most automatable function — monitoring, detection and tracking transfer to sensors + AI first; it's already happening. Established
- Physical maintenance and repair are the least automatable — the engine room needs hands; no near-term robotics replaces an engineer in a space not designed for machines. Maritime-GPT analysis
- Judgement under ambiguity resists automation — COLREGS edge cases, negotiation, non-standard situations are exactly where the AI evidence is weakest. Established
- The new role isn't 'no seafarer' — it's Remote Operations Centre operator, a relocated maritime professional the MASS Code now names. Established
- Expert consensus on future skills ranks Operational, Digital and Maritime competencies — the seafarer of 2035 is differently skilled, not absent. Research demonstration
Function by function — what actually automates
| Function | Automation outlook | Evidence | What it means for the seafarer |
|---|---|---|---|
| Watchkeeping / lookout | High — automating now | Research demonstration | Moves to supervision of automated watch; fewer eyes, more screens |
| Collision avoidance (routine) | Medium — assistive | Research demonstration | AI proposes, officer decides — support, not replacement |
| Voyage planning / optimization | High — largely automated | Industry deployment | Planner becomes reviewer of optimized options |
| Machinery monitoring / diagnostics | High — proven | Field trial | Reactive repair → planned maintenance on early warning |
| Physical maintenance / repair | Low — hard to automate | Maritime-GPT analysis | The engine room keeps its humans longest |
| Judgement / command / leadership | Low — resists automation | Maritime-GPT analysis | The master's accountability role stays human by regulation |
| Remote operations (new) | Emerging role | Established | The ROC operator — a seafarer's skills, ashore |
The pattern is consistent: information work automates, physical and judgement work doesn't. Monitoring, tracking, planning and diagnostics are already transferring to AI; hands-on repair, non-routine judgement and accountable command are not — and the last is fixed in regulation, not just technology. Maritime-GPT analysis
The role that replaces — not removes
The MASS Code didn't abolish the seafarer; it invented a new address for one. The Remote Operations Centre is now a regulated component of autonomous shipping — a shore station where maritime professionals do watchkeeping, monitoring and intervention for vessels that may have reduced or no crew aboard. The skills don't disappear; they relocate — same situation awareness, same COLREGS judgement, different chair. Established
And the expert data says the role gets broader, not narrower: the 174-expert TOPSIS study ranks Operational Skills, Digital Skills and Maritime Competency as the top MASS requirements — a hybrid mariner-technologist. The PRISMA review of maritime education reaches the same place: MET needs updating for autonomy, not abandoning, and there's no standardised autonomy-training framework yet to fill the gap. Research demonstration
The honest constraint nobody quotes
The people being automated have a vote, and they're sceptical. The 1,009-officer survey found trust in autonomy is low — lowest among the most professionally committed officers — and respondents demand human presence, oversight and control, plus reliability and redundancy before they'll trust it. Research demonstration
That's not Luddism; it's the workforce identifying the same failure modes the safety literature does (deskilling, out-of-the-loop degradation, automation bias). An industry that automates against its most skilled officers' judgement will struggle to recruit the ones it still needs — the sceptics are the talent. Maritime-GPT analysis
What we know / don't / being tested
What we know
- Information functions (watchkeeping, planning, diagnostics) are automating first.
- Regulation keeps human command and creates the ROC role.
- Future skills are hybrid maritime + digital, per expert ranking.
What we don't know
- Pace — how fast autonomy moves from pilot to material crew reduction.
- The ROC staffing model — how many officers ashore oversee how many ships.
- Whether training pipelines produce hybrid mariner-technologists fast enough.
Being tested
- Reduced-manning and remote-operation pilots under flag supervision.
- New MET curricula for autonomy-era competencies.
- The trust/adoption question — whether officers will work the new way.
Maritime-GPT analysis
AI won't replace the seafarer; it will replace the seafarer who only does what AI does. The functions moving first are the routine information tasks — the ones that already bored officers. What stays is everything AI can't do: physical presence, non-routine judgement, leadership, and legal accountability. The career risk isn't automation taking the job — it's automation taking the interesting parts while a poorly designed system leaves the human the supervision shift nobody wants. The future seafarer is a supervisor, diagnostician and remote operator; the sooner training matches that, the smoother the transition. Maritime-GPT analysis
Open research problems
- ROC competency standards — what certifies a shore operator the way STCW certifies a ship officer.
- Skill-preserving job design — keeping offshore competence alive when routine tasks automate.
- Crewing economics — the real reduction-vs-relocation balance across vessel types.
- Recruitment — whether "AI-assisted seafaring" attracts or repels the next generation.
Evidence scorecard
| Dimension | Rating | Basis |
|---|---|---|
| Evidence strength | High | Regulatory record + multiple peer-reviewed workforce studies. |
| Certainty of direction | High | Information tasks automate, physical/judgement tasks persist — consistent. |
| Certainty of pace | Unclear | Timeline depends on regulation, economics and acceptance — not just tech. |
| Confidence in conclusions | High | Claims from official instruments and named studies. |
Related tools & reading
Sources
- Autonomous shipping — MASS regulatory scoping exercise and degrees of autonomy — International Maritime OrganizationPrimary — regulatory
The foundational constraint: STCW's applicability is defined by seafarers being aboard; autonomy degrees 1–4 set the frame for what 'replace' even means. - IMO adopts first global Code for autonomous ships (MASS Code) — International Maritime OrganizationPrimary — regulation
The 2026 code keeps the master responsible at all times and introduces the Remote Operations Centre — the new role, not the removed one. - The human element in autonomous shipping: skills and competency requirements (fuzzy TOPSIS, 174 experts) — WMU Journal of Maritime Affairs (2025)Peer-reviewed
Expert ranking of MASS-era skills: Operational Skills, Digital Skills and Maritime Competency top the list — the job changes, it doesn't vanish. - Challenges for the education and training of seafarers in the context of autonomous shipping (PRISMA review) — Applied Sciences 14(8):3173Peer-reviewed review
Systematic review of MET for MASS — the skills-transition evidence base and the absence of standardised autonomy training frameworks. - Trust in autonomous vessels — mixed-method analysis of 1,009 bridge officers — WMU Journal of Maritime Affairs (2025)Peer-reviewed
The workforce reality: low trust, strong demand for human presence and oversight — the people being 'replaced' get a say.
Regulatory statements from IMO instruments and scoping exercise records; workforce findings from named peer-reviewed studies. Accessed 5 October 2026. Editorial analysis, not career or legal advice.