Gains:
- Ability to establish situation awareness with radar, AIS, ECDIS and eye cross-checking and use artificial intelligence in target prioritization and briefing
- Ability to understand that the collision avoidance maneuver belongs to the officer of the watch within the framework of COLREG and good maritime practice and cannot be transferred to the system.
- Recognizing the dangers of over-reliance on automation (automation complacency) and dependence on a single source, and being able to prioritize safety when advice conflicts with practice.
The bridge is the brain of the ship. Radar, ECDIS, AIS (Automatic Identification System - the system where ships broadcast their identity and movement information), radio and lookout all together create situational awareness - understanding what everything in the environment is and what it will do. With heavy traffic, narrow waters and poor visibility, decision-making pressure is high. AI and intelligent systems can provide decision support on the bridge; However, collision avoidance maneuvers and navigational decisions belong to the officer of the watch and the captain within the framework of COLREG. In this unit, you will learn how to safely use AI on the bridge, the pitfalls of over-reliance on automation (automation complacency), and human-machine collaboration.
COLREG and conflict avoidance: it's up to the person to decide
COLREG (International Rules for Preventing Collision at Sea) determines who will give way and how to maneuver when two ships meet. Key concepts:
- CPA/TCPA (Closest Point of Approach / Time to CPA — the closest passage distance of two ships and the time remaining until that moment): measures the risk of collision.
- Give-way / stand-on (ship giving way / maintaining course): Who will maneuver according to COLREG.
- Early and clear maneuver: COLREG sees the risk early and demands a clear and understandable maneuver.
Modern radar/ARPA (automatic target tracking) systems calculate CPA/TCPA; AI-powered systems can even produce risk rankings and maneuver recommendations. But the decision to maneuver — how many degrees, when, which direction — rests with the officer of the watch and is based on COLREG, good seamanship and the immediate context.
Caution: Even if an AI or automation says "safe pass", the responsibility for the collision lies with the officer who performed (or did not perform) the maneuver. The COLREG decision is not transferable. When system recommendation conflicts with good maritime practice, practice and safety come first.
The danger of over-reliance on automation
As systems operate reliably, human attention may decrease; This is called automation complacency (relying too much on automation and giving up control). There are known consequences of this at sea: trusting the screens and not looking out of the window, missing the small boat that is not visible on the AIS, mistaking the ECDIS warning for routine.
Safe bridge principles:
- Cross-check: Radar, AIS, ECDIS and eye sight must confirm each other; One does not replace the other. The boat without AIS is radar/visible.
- Automation is an input, not a decision: AI recommendation is one of four or five sources of information.
- The human is in the loop: In critical decisions, the human thinks actively; Just because the system says "everything is fine" does not remove the control.
Tip: When using an AI-enabled bridge vehicle, ask yourself regularly: “What independent source did I confirm this recommendation with?” Ask the question. Radar target and AIS and eye, are all three saying the same thing? If there is dissociation, take the most dangerous interpretation as basis.
Safe contributions of AI on the bridge
- Situation summary: In heavy traffic, list the targets in order of risk and a summary such as "these are the three highest risk targets" (to direct the officer's attention).
- Checklists: Departure, arrival, narrow water crossing, restricted visibility procedure checklists.
- Briefing and handover: Draft situation briefing at shift change.
- Training and exercise: Scenario generation, COLREG status discussion.
- Information access: Port information, procedure summary, signal/light meaning reminder (confirmation with official source).
In all of them, AI directs attention and prepares; The officer makes the maneuver and navigation decisions.
three mini cases
Case 1 — Attention redirection. There are many targets in a busy strait crossing. The decision support system highlights the three riskiest targets according to CPA/TCPA. The officer of the watch focuses his attention on this summary, but confirms each target with radar and visual and plans his own maneuver according to COLREG. The system prioritized; The officer decided.
Case 2 — Invisible boat. While a ship is relying on AIS traffic and proceeding, a small fishing vessel without AIS enters the route. Lookout and radar boat sees; The team maneuvers in a timely manner. If only AIS/automation had been relied upon the boat would have been missed. Lesson: the system works with incomplete data; eye and radar are indispensable.
Case 3 — Conflict between recommendation and practice. A decision support tool recommends “quick pass” in a narrow channel; But the officer sees the wave produced by the oncoming large ship and the narrowness of the channel and chooses to slow down and wait. Good maritime practice trumps the suggestion of automation. Lesson: the last word lies with the person who sees the context.
Four copyable templates
1) Shift handover briefing:
Your role: bridge assistant. I will give you the current navigation status (location, route, speed, traffic, weather, caution issues). Task: write a 10-line draft handover briefing for the incoming officer of the watch. Only use the information I provide; Don't suggest decisions/manoeuvres, just summarize the situation. Mark the uncertain.
2) Pre-migration checklist:
Draft a bridge readiness checklist for [narrow water crossing / harbor approach / limited visibility]: equipment check, crew duties, communications, emergency preparedness. Note: this is a general reminder; It will be completed with ship procedure and captain's instructions.
3) Traffic situation summary (not decision):
I will give you the radar/AIS target list (distance, bearing, CPA, TCPA). Task: rank targets in order of risk and summarize those with the highest risk. I will make the COLREG maneuver decision and confirm each target with radar + eyes; You just prioritize, DO NOT order maneuvers.
4) COLREG training scenario:
Generate a COLREG encounter scenario (e.g. narrow channel encounter) and add 3 discussion questions. The purpose is to discuss the logic of the rule. Giving precise maneuver "prescription"; Emphasize that the answer may vary depending on the ship and context.
Weak prompt / Strong prompt
Weak prompt:
How should I avoid this ship, what maneuver should I do?
It is wrong to ask AI to make maneuver decisions; Doesn't know context, COLREG role, other traffic.
Powerful prompt:
Your role: bridge decision support assistant. I'll give you my radar/AIS target list (distance, bearing, CPA, TCPA, ship type) and my own heading/speed. Task: rank targets by collision risk and summarize the most critical; Write down what information I need to confirm for each. I will make the COLREG maneuver decision and time; Don't order maneuvers, just direct my attention.
It is safe to keep the AI as a prioritizer and reminder, leaving the decision to the officer.
Bridge: role distribution
Quest
AI/automation
officer/captain
Goal prioritization
Risk ranking
Confirmation + decision
CPA/TCPA
Account (ARPA)
Comment, maneuver
COLREG maneuver
(Out of decision)
Officer decision
situation briefing
draft
verification
checklists
draft
Completion with procedure
Small/non-AIS boat
may miss
eye + radar
Common mistakes
- Leaving the maneuver decision to the system. The COLREG decision lies with the officer; suggestion is input, not command.
- Just relying on AIS/display. Boats without AIS are visible to radar and sight; automation works with incomplete data.
- Automation complacency. Even though the system works reliably, auditing and cross-checking are not allowed.
- To put the proposal ahead of good maritime practice. Context and safety may override system recommendation.
- Ignoring separation. If radar, AIS and eyes say differently, the most dangerous interpretation is taken as basis.
In summary
Situational awareness on the bridge is achieved through radar, AIS, ECDIS and eye cross-checking. AI and decision support systems enhance officer attention and preparedness through objective prioritization, briefing, and checklists. But the maneuver to avoid collision is the responsibility of the officer of the watch and the captain within the framework of COLREG and good maritime practice; Over-reliance on automation and dependence on a single source is dangerous. The system suggests; man sees, confirms and decides.
Application task
Set up a traffic scenario: your own route and 5-6 destinations (distance, bearing, CPA, TCPA). Have AI prioritize with the "Traffic situation summary" template. Don't make one of them a "small boat without AIS" and put it on the list; Write about why the AI cannot see it and why eye/radar confirmation is essential. Add your own COLREG assessment for each critical objective.
checklist
- [ ] As the officer, I made the maneuver and COLREG decision; I did not leave it to the system.
- [ ] I cross-checked the radar, AIS and eye; I did not trust a single source.
- [ ] I maintained eye and radar confirmation for non-AIS/small vessels.
- [ ] When the automation recommendation conflicted with good maritime practice, I relied on practice.
- [ ] I kept active control against automation complacency.