Civil Law and Autonomous Shipping Navigation Failure Claims in Europe

1. Introduction

Autonomous shipping navigation refers to the use of automated or AI-supported systems to navigate ships with reduced or no direct human control. These systems may perform:

route planning;

collision avoidance;

steering;

speed control;

weather routing;

radar and sensor interpretation;

electronic chart analysis;

berth approach;

obstacle detection;

autonomous manoeuvring;

communication with shore-control centres.

A navigation failure can occur when an autonomous system:

fails to detect another vessel;

incorrectly interprets radar or AIS information;

selects an unsafe route;

fails to react to weather;

gives an incorrect collision-avoidance instruction;

loses communication with a remote-control centre;

suffers a software or sensor failure;

follows outdated navigational data; or

improperly overrides human intervention.

European law does not yet have a large body of reported cases specifically deciding liability for fully autonomous vessels. Current maritime liability rules were largely developed for conventional ships operated by masters and crews. Recent scholarship likewise identifies uncertainty over how collision rules should apply to autonomous vessels and how responsibility should be divided between shipowners, operators, remote-control centres and technology providers. (Sage Journals)

Accordingly, the existing maritime cases below are primarily analogical authorities, although their principles concerning seaworthiness, navigation, collision, causation and carrier liability can be highly relevant.

2. Meaning of an Autonomous Navigation Failure

An autonomous navigation failure may occur at several technological levels.

1. Sensor failure

The vessel incorrectly receives information from:

radar;

lidar;

cameras;

GPS;

AIS;

depth sensors.

2. Data failure

The system receives:

outdated charts;

incorrect coordinates;

inaccurate weather information;

corrupted navigational data.

3. Algorithmic failure

The AI incorrectly calculates:

collision risk;

safe distance;

route;

speed;

turning angle.

4. Communication failure

The ship loses connection with:

shore-control centre;

satellite system;

navigation network;

other vessels.

5. Human-supervision failure

The autonomous system identifies a danger but the remote operator fails to intervene.

6. Cybersecurity failure

A malicious actor alters:

navigation data;

GPS;

route instructions;

steering commands.

3. Main Legal Question

The central question is:

Who is legally responsible when an autonomous vessel causes damage because its navigation system fails?

Potentially relevant parties include:

shipowner;

operator;

master or remote operator;

charterer;

autonomous-navigation-system manufacturer;

software developer;

sensor manufacturer;

data provider;

remote-control-centre operator;

classification society;

maintenance provider.

The answer depends upon the applicable maritime convention, national law, contract, regulatory framework and facts.

4. Applicable European Legal Framework

A. Maritime collision law

Collision disputes may engage:

COLREGs;

national maritime law;

1910 Collision Convention where applicable;

limitation-of-liability rules;

insurance arrangements.

The traditional system focuses on the fault of vessels and persons involved in navigation.

Autonomous vessels create an attribution problem:

If there is no human actively steering the vessel, whose fault is the vessel's navigational error?

5. COLREGs and Autonomous Navigation

The International Regulations for Preventing Collisions at Sea remain particularly important.

Autonomous systems must effectively perform functions corresponding to traditional navigational duties, including:

maintaining proper lookout;

proceeding at safe speed;

assessing collision risk;

taking appropriate avoiding action;

observing right-of-way rules;

conducting appropriate manoeuvres.

The technological question becomes:

Can an autonomous system satisfy a rule that traditionally assumes human judgment?

For example, a navigation AI may technically maintain a radar watch, but if its algorithm fails to recognise a small fishing vessel, a collision claim may arise.

6. Seaworthiness

One of the most important civil-law concepts is seaworthiness.

A vessel should be reasonably fit for:

the intended voyage;

navigation;

foreseeable conditions;

carrying its cargo safely.

For an autonomous vessel, seaworthiness may increasingly include:

reliable navigation software;

properly calibrated sensors;

cybersecurity;

reliable communications;

adequate backup systems;

properly tested AI;

appropriate human supervision.

Thus:

Digital seaworthiness may become as important as traditional physical seaworthiness.

7. Case Law

Case 1: The CMA CGM Libra [2021] UKSC 51

This is one of the most important modern European maritime authorities on seaworthiness.

The case concerned defective passage planning and whether the vessel was unseaworthy because of an inadequate passage plan.

Principle

The UK Supreme Court treated proper passage planning as part of the vessel's seaworthiness obligations.

Relevance to autonomous navigation

An autonomous ship may generate its own passage plan.

Suppose its navigation algorithm:

selects an unsafe route;

fails to account for a known hazard;

uses inadequate chart information.

The CMA CGM Libra reasoning provides a powerful analogy:

Navigation planning is not merely an operational detail; deficiencies in navigation preparation can potentially constitute unseaworthiness.

Therefore, an autonomous vessel's algorithmic route-planning system could become part of the seaworthiness analysis.

8. Case 2: The Eurasian Dream [2002] 1 Lloyd's Rep 719

This English Admiralty case concerned the standard of seaworthiness and the competence required for proper operation of a vessel.

Principle

Seaworthiness extends beyond the physical condition of the ship and can involve the competence and ability necessary for the vessel's safe operation.

Application to autonomous vessels

The relevant question could become:

Was the autonomous-navigation system reasonably fit and properly configured for the voyage?

This could include:

software validation;

sensor integration;

navigation-data quality;

system testing;

remote-operator competence.

Therefore, the case provides an important analogy for technological seaworthiness.

9. Case 3: The Lady Gwendolen [1965] 1 WLR 1482

This classic collision case concerned navigational fault and the application of collision principles.

Principle

Liability in maritime collision can depend upon whether a vessel's navigational conduct departed from the applicable standard.

Autonomous-navigation relevance

The traditional question:

“Did the navigator act reasonably?”

may need to become:

“Was the autonomous navigation system reasonably designed, operated, supervised and maintained?”

An autonomous vessel should not necessarily receive a lower standard merely because the navigational decision was made by software.

10. Case 4: The Bow Spring [2004] EWHC 1363 (Admlty)

This Admiralty case concerned navigational conduct and collision liability.

Principle

Maritime courts examine the actual circumstances surrounding navigation, including:

visibility;

vessel movements;

available information;

navigational decisions;

compliance with collision rules.

Relevance

For an autonomous ship, the court could reconstruct:

sensor data → algorithmic assessment → navigation decision → manoeuvre → collision.

The case therefore illustrates why detailed reconstruction of navigational events is essential.

11. Case 5: Monford Management Ltd v Afina Navigation Ltd [2026] EWCA Civ 251

This is a particularly recent collision authority.

The case concerned the collision between KIVELI and AFINA I off the south coast of Greece. The Court of Appeal dealt with issues concerning the Collision Regulations and apportionment of fault. The trial court had found KIVELI 80% at fault and AFINA I 20% at fault. (Courts and Tribunals Judiciary)

Relevance to autonomous vessels

The case demonstrates the continuing importance of:

COLREG interpretation;

determining when collision risk arose;

navigational decisions;

causation;

apportionment of fault.

For autonomous vessels, similar analysis could be applied to the algorithm's navigational decisions.

The technological issue would then be:

Which system failure or operational decision produced the vessel's COLREG breach?

12. Case 6: The Ocean Victory [2017] UKSC 35

This major Supreme Court maritime case concerned charterparty obligations, port safety and the interaction between contractual maritime obligations and physical conditions.

Principle

Maritime contractual liability depends heavily upon the precise contractual allocation of risk and the circumstances of the voyage.

Relevance to autonomous vessels

Autonomous shipping contracts may allocate responsibility among:

shipowner;

charterer;

autonomous-system operator;

remote-navigation provider.

Therefore, when navigation fails, the court must examine the relevant contract rather than assuming that every loss belongs to the shipowner.

13. Case 7: Volcafe Ltd v Compania Sud Americana de Vapores SA [2018] UKSC 61

This Supreme Court case concerned cargo liability and the carrier's obligations regarding care of cargo.

Principle

The case is important for the evidential and burden-of-proof dimensions of maritime liability.

Relevance

Autonomous navigation failures can cause cargo damage even when the vessel itself suffers little or no physical damage.

For example:

Autonomous route-planning error → severe weather → cargo damage.

The legal analysis may then concern:

carrier's obligations;

seaworthiness;

causation;

evidence;

contractual exclusions.

14. Case 8: MH v Costa Crociere, C-629/24, CJEU, 4 June 2026

This is a particularly recent CJEU maritime-liability judgment.

The case concerned passenger injury during a cruise and the application of Regulation (EC) No 392/2009 and the Athens Convention framework.

The CJEU examined the scope of carrier liability, insurance and compensation limits. The judgment expressly addresses shipping incidents including collision, stranding and defects in a ship, including malfunction or failure affecting propulsion, steering and safe navigation. (Eur-Lex)

Relevance to autonomous vessels

This is highly relevant conceptually because autonomous navigation failures may involve:

steering;

propulsion;

safe navigation;

system malfunction.

If an autonomous system fails and passengers are injured, the statutory maritime carrier-liability framework may remain applicable.

The fact that the ship used autonomous technology does not automatically remove it from the maritime liability regime.

15. Case 9: Wallentin-Hermann v Alitalia, C-549/07

Although concerning air transport rather than shipping, this CJEU case is useful by analogy for technological transport failures.

Principle

Technical problems do not automatically qualify as extraordinary circumstances relieving a carrier from responsibility.

Relevance

An autonomous ship operator should not automatically argue:

“The software malfunction was unexpected, therefore there is no liability.”

The court may need to examine:

whether the malfunction was foreseeable;

whether reasonable maintenance existed;

whether backup systems existed;

whether the operator could reasonably control the risk.

The analogy must be treated cautiously because the case concerns aviation.

16. Case 10: McDonagh v Ryanair, C-12/11

Again, this is an aviation rather than maritime case.

Principle

Transport operators may have continuing obligations even during extraordinary disruption.

Relevance

The case illustrates a broader European transport principle:

Technological or operational disruption does not automatically extinguish passenger-protection obligations.

For autonomous ships carrying passengers, the precise maritime rules would govern, but the principle is useful by analogy.

17. Collision Between Autonomous and Conventional Vessel

Consider:

Autonomous Vessel A fails to detect Vessel B and collides with it.

Potential causes include:

defective radar;

AI recognition error;

incorrect COLREG interpretation;

outdated electronic charts;

communication failure.

The court may ask:

Step 1

Did the autonomous vessel breach a navigation rule?

Step 2

Was the breach caused by:

software;

sensor;

operator;

maintenance;

data?

Step 3

Who legally controlled the vessel?

Step 4

Was the ship seaworthy?

Step 5

Was the defect discoverable before departure?

Step 6

Did the other vessel also contribute to the collision?

Step 7

How should liability be apportioned?

18. Software Defect and Seaworthiness

Suppose an autonomous ship leaves Rotterdam.

Before departure, the navigation software contains a known defect that causes it to misinterpret AIS signals.

During the voyage:

AIS signal → incorrect algorithmic interpretation → incorrect manoeuvre → collision.

The claimant could argue:

the vessel was unseaworthy;

the software was defective;

the owner failed to test it;

the navigation system was improperly maintained.

The shipowner might argue:

the defect was undiscoverable;

the system complied with applicable certification;

the collision resulted from another vessel;

the software provider was contractually responsible.

This makes pre-voyage testing and documentation extremely important.

19. Remote-Control Centre Liability

Autonomous ships may be supervised from land.

A remote operator may:

monitor several ships;

intervene during emergencies;

approve route changes;

respond to sensor warnings.

Suppose:

Autonomous system detects a collision risk → sends warning to remote centre → operator ignores warning → collision.

Potential responsibility could involve:

shipowner;

remote operator;

remote-control centre;

software provider.

The factual question becomes whether the operator had a legally meaningful opportunity and duty to intervene.

20. Manufacturer Liability

The navigation-system manufacturer may potentially face liability where:

software was defective;

sensors were defective;

system specifications were inadequate;

safety warnings were insufficient;

cybersecurity was inadequate;

updates introduced dangerous defects.

However, a manufacturer should not automatically be responsible for every accident.

The claimant may need to establish:

defect → causation → damage.

Contractual warranties and applicable product-liability rules will also matter.

21. Cyberattack

Cybersecurity is particularly important for autonomous vessels.

Imagine:

Hacker alters GPS coordinates → autonomous vessel believes it is on a safe route → vessel enters dangerous waters → grounding occurs.

Possible claims could involve:

shipowner;

cybersecurity contractor;

navigation-system provider;

software developer.

Questions include:

Was cybersecurity adequate?

Was the attack foreseeable?

Were reasonable security controls implemented?

Was there redundant navigation?

Was human intervention available?

A cyberattack may provide a defence in some circumstances, but it does not automatically eliminate liability.

22. Grounding

Autonomous navigation can also cause grounding.

Possible causes:

incorrect depth data;

chart error;

GPS failure;

sensor malfunction;

algorithmic route selection;

failure to account for tides.

The CMA CGM Libra principle becomes particularly useful.

If the autonomous system generates a defective passage plan, the court may examine whether that defect amounts to a seaworthiness problem.

23. Cargo Damage

An autonomous navigation failure may cause:

collision;

grounding;

heavy rolling;

container loss;

water ingress;

temperature-control failure.

The cargo owner may sue the carrier under the applicable carriage regime.

The carrier may need to establish compliance with its obligations concerning:

seaworthiness;

navigation;

cargo care;

reasonable precautions.

24. Passenger Injury

Autonomous cruise ships or ferries raise another layer of liability.

Potential incidents include:

collision;

sudden manoeuvre;

grounding;

failure to avoid an obstacle;

docking accident.

The recent Costa Crociere, C-629/24 judgment is especially relevant because EU maritime passenger liability rules expressly address defects or failures affecting safe navigation. (Eur-Lex)

25. Evidence

Autonomous navigation litigation will depend heavily on technical evidence.

Important evidence may include:

radar recordings;

AIS records;

GPS records;

electronic chart data;

sensor outputs;

algorithm logs;

source-code documentation;

software version;

system updates;

cybersecurity records;

remote-control communications;

weather data;

voyage plan;

maintenance records;

classification records;

emergency-system logs.

The court may need to reconstruct:

What the vessel knew → what the algorithm calculated → what action it ordered → whether the action complied with navigation rules → what caused the damage.

26. Black-Box Problem

A particularly difficult problem is algorithmic opacity.

Suppose an AI system decides:

“Turn 20° starboard.”

But nobody can easily explain why.

After a collision, the claimant may ask:

What data did the system use?

Which object did it identify?

What collision probability did it calculate?

Which COLREG rule did it apply?

Why did it select that manoeuvre?

This makes audit logs and explainability increasingly important.

27. Causation

Autonomous navigation disputes require careful causal analysis.

Example:

Sensor failure

↓

Incorrect AI perception

↓

Incorrect route

↓

Collision

↓

Cargo damage

The claimant must establish that the alleged failure actually contributed to the damage.

Other possible causes include:

another vessel's negligence;

bad weather;

mechanical failure;

incorrect chart;

human intervention;

cyberattack.

28. Contributory Fault

Suppose an autonomous vessel makes an incorrect manoeuvre, but the other vessel also:

fails to maintain lookout;

travels too fast;

fails to take avoiding action;

incorrectly interprets the collision situation.

Liability may therefore be apportioned.

The recent Monford Management v Afina Navigation judgment demonstrates the continuing importance of analysing the respective navigational faults and apportioning responsibility under collision rules. (BAILII)

29. Force Majeure

Shipowners may argue that navigation failure resulted from:

extreme weather;

cyberattack;

satellite outage;

unexpected software failure;

loss of communications.

But a force-majeure argument depends upon the relevant contract and applicable law.

The court may ask:

Was the event unforeseeable?

Was it beyond control?

Could reasonable precautions have prevented it?

Were backup systems available?

Was the failure caused by inadequate maintenance?

30. Autonomous Vessel vs Autonomous Vessel

A particularly difficult future situation is:

Autonomous Vessel A collides with Autonomous Vessel B.

There may be no human decision-maker physically aboard either vessel.

The investigation may instead examine:

Algorithm A;

Algorithm B;

sensor systems;

software versions;

remote operators;

shipowners;

technology providers.

The fundamental civil-law question becomes:

How should traditional concepts of navigational fault be attributed when navigation is performed by software?

Current European law does not yet provide a fully settled answer for every such scenario. Recent research expressly identifies this uncertainty. (Sage Journals)

31. Traditional Maritime Fault vs Algorithmic Fault

The legal principles may remain similar, but attribution becomes more complicated.

32. Digital Seaworthiness

A future court may increasingly need to examine whether a vessel was digitally seaworthy.

This could involve:

Software

Was it properly developed and tested?

Sensors

Were they functioning correctly?

Data

Was navigational information accurate?

Cybersecurity

Was the system protected?

Redundancy

Was there a backup system?

Human supervision

Was qualified intervention available?

Updates

Was the software properly maintained?

Thus, seaworthiness may increasingly include both:

physical seaworthiness + digital seaworthiness.

33. Contractual Allocation of Risk

Autonomous shipping contracts should address:

navigation-system responsibility;

software updates;

sensor maintenance;

cybersecurity;

remote-control responsibilities;

emergency intervention;

data ownership;

system downtime;

liability caps;

indemnities;

insurance;

audit rights;

incident reporting;

evidence preservation.

This becomes especially important where several companies participate in the autonomous-navigation ecosystem.

34. Insurance

Autonomous vessels create questions concerning:

hull insurance;

cargo insurance;

protection and indemnity insurance;

cyber insurance;

product liability insurance;

professional liability insurance.

An insurer may investigate:

Was the loss caused by ordinary maritime risk, human negligence, software defect, cyberattack or inadequate maintenance?

35. Six Key Cases for Examination

Additional useful analogical authorities include The Ocean Victory [2017] UKSC 35 and Volcafe v CSAV [2018] UKSC 61.

36. Important Qualification About the Case Law

The above cases should not be described as cases in which courts directly decided that an AI autonomous vessel was liable.

There is currently a relatively small body of direct European judicial authority concerning fully autonomous maritime navigation. The Monford Management litigation, for example, concerned conventional vessels, although its collision principles are relevant to autonomous shipping. (BAILII)

Similarly, CMA CGM Libra concerned a conventional vessel and passage planning, but its seaworthiness reasoning provides an important analogy for autonomous route-planning systems.

This distinction is important for academically accurate legal writing.

37. Practical Liability Test

A court dealing with an autonomous-navigation failure could analyse the dispute in this order:

Step 1 — Identify the accident

Was it:

collision;

grounding;

cargo loss;

passenger injury;

environmental damage?

Step 2 — Identify the navigation failure

Was it caused by:

software;

sensor;

data;

communications;

human supervision;

cyberattack?

Step 3 — Examine seaworthiness

Was the vessel reasonably fit for the voyage, including its autonomous-navigation system?

Step 4 — Examine COLREG compliance

Did the vessel comply with applicable collision-avoidance requirements?

Step 5 — Identify the responsible actor

Was responsibility with:

owner;

operator;

remote centre;

manufacturer;

software provider?

Step 6 — Establish causation

Did the technological failure cause the accident?

Step 7 — Examine contributory fault

Did another vessel or party contribute?

Step 8 — Calculate damages

Possible damages include:

vessel damage;

cargo loss;

personal injury;

environmental loss;

delay;

economic loss.

38. Hypothetical Example

Facts

An autonomous container ship travels from Rotterdam to Greece.

Its navigation AI receives an incorrect electronic-chart update.

The system calculates that a particular channel is sufficiently deep.

The ship enters the channel, grounds and damages its cargo.

Possible claims

Cargo owner → carrier

Possible breach of carriage obligations.

Shipowner → software provider

Possible contractual/product claim if the software caused the error.

Shipowner → data provider

Potential claim if inaccurate chart data was supplied contrary to contractual obligations.

Seaworthiness question

Was the ship reasonably seaworthy when it began the voyage?

Algorithm question

Was the AI properly tested to detect inconsistent chart data?

Causation

Did the chart error actually cause the grounding?

Defence

Was the defective chart information reasonably discoverable?

This demonstrates how traditional maritime doctrines can accommodate autonomous-technology disputes without creating an entirely new liability system.

39. Key Legal Principles

Autonomous vessels do not automatically possess separate legal personality.

The shipowner or operator normally remains an important potential bearer of liability.

Autonomous navigation does not eliminate COLREG obligations.

Software may become part of the seaworthiness analysis.

Digital seaworthiness may include software, sensors, data and cybersecurity.

A defective autonomous route may potentially raise the same basic concerns as defective human passage planning.

Technology providers may bear separate contractual or product-related responsibility.

Remote-control operators may become relevant where human intervention is expected.

Collision liability may still require analysis of fault and causation.

Contributory fault can result in apportionment between vessels.

Electronic logs may become the maritime equivalent of traditional navigational records.

Cyberattack does not automatically eliminate liability.

Passenger maritime-liability regimes can apply to accidents involving defects affecting safe navigation.

Contractual risk allocation becomes particularly important where several technology providers are involved.

Fully autonomous vessel liability remains an evolving area of European maritime law.

40. Exam-Ready Formula

For an autonomous shipping navigation failure, use:

Autonomous vessel → navigation system → sensor/data/software failure → COLREG/seaworthiness issue → responsible actor → causation → contributory fault → applicable maritime convention/contract → damages → limitation/defence → remedy.

41. Conclusion

Autonomous shipping navigation failure claims represent a technological extension of traditional European maritime civil liability rather than a completely separate legal category. The principal issues remain seaworthiness, navigational fault, collision rules, contractual obligations, causation, cargo/passenger liability and allocation of loss, but autonomous technology adds difficult questions concerning software, sensors, data, cybersecurity and remote supervision.

The CMA CGM Libra is particularly useful for analysing autonomous passage planning and digital seaworthiness; The Eurasian Dream assists with the broader concept of seaworthiness; traditional collision cases such as The Lady Gwendolen and The Bow Spring help analyse navigational fault; Monford Management v Afina Navigation provides a very recent illustration of COLREG interpretation and fault apportionment; and the CJEU's MH v Costa Crociere (C-629/24) provides a current EU maritime-liability authority concerning accidents and defects affecting safe navigation. (Eur-Lex)

The principal future legal question is therefore not simply “Who programmed the ship?”, but:

Who had the legal duty to ensure that the autonomous vessel was seaworthy, properly navigated, adequately supervised and reasonably protected against foreseeable technological failures?

That question will determine whether liability ultimately falls on the shipowner, operator, remote-control centre, software developer, sensor/data provider, another vessel, or a combination of parties.