Civil Law and Autonomous Space Systems Liability Claims in Europe

1. Introduction

Autonomous space systems are spacecraft, satellites, orbital vehicles, space robots, autonomous docking systems, lunar/planetary vehicles, autonomous navigation systems, satellite constellations, and other space systems capable of making operational decisions without continuous human intervention.

Examples include:

autonomous satellites changing orbit;

spacecraft performing collision avoidance automatically;

autonomous docking and rendezvous;

robotic lunar or planetary vehicles;

AI-controlled satellite navigation;

autonomous space-station systems;

satellite constellations automatically reallocating network capacity;

autonomous debris-avoidance systems;

software-controlled spacecraft that detect and correct faults.

European law does not generally give an autonomous space system an independent legal personality. Therefore, when an autonomous spacecraft causes injury, property damage, satellite collision, service interruption, or economic loss, liability normally has to be attributed to a human or corporate legal actor such as the manufacturer, operator, launch provider, software developer, service provider, owner, or controlling organisation.

There is currently very limited direct European case law concerning an autonomous spacecraft itself causing damage. Consequently, the strongest legal analysis combines international space-liability rules with European product-liability, contract, tort/delict and technology-liability jurisprudence. EU institutional material itself has recognised that autonomous systems create difficulties in identifying the source of damage and allocating liability. (Eur-Lex)

2. Meaning of Autonomous Space-System Liability

Autonomous space-system liability means the civil or international legal responsibility arising when an autonomous space system:

malfunctions;

makes an incorrect autonomous decision;

collides with another space object;

damages another satellite;

causes injury to persons;

damages property on Earth;

loses control and creates orbital debris;

provides incorrect navigation or communication services;

suffers a cybersecurity attack;

causes commercial or economic losses.

The central legal question is:

Who is legally responsible when the immediate harmful action was taken by an autonomous machine rather than directly by a human?

The answer generally depends upon attribution, applicable liability regime, defect or fault, causation and damage.

3. Applicable European Legal Framework

A. International Space Law

European space activities operate within the international space-law framework, particularly:

Outer Space Treaty 1967;

Liability Convention 1972;

Registration Convention 1975;

Rescue Agreement 1968;

Moon Agreement 1979, where applicable.

The Liability Convention is particularly important.

It distinguishes between:

1. Damage on Earth or to aircraft in flight

Liability of the launching State is generally based on absolute liability.

Therefore, proving negligence may not be necessary at the international level.

2. Damage caused elsewhere in outer space

For damage to another space object or persons/property aboard it, liability generally depends upon fault.

This creates a major distinction:

4. EU and European Civil-Law Rules

The international space regime does not necessarily exhaust private-law liability.

A claim may additionally involve:

Contract law

For example:

satellite-service contracts;

launch contracts;

insurance agreements;

satellite-operation agreements;

software-maintenance agreements;

telecommunications contracts.

Tort/delict law

Possible claims include:

negligence;

breach of statutory duty;

unsafe operation;

failure to maintain;

failure to warn;

inadequate cybersecurity.

Product liability

An autonomous spacecraft may contain:

hardware;

sensors;

processors;

software;

navigation systems;

propulsion systems;

AI systems.

European product-liability jurisprudence is therefore highly relevant.

Data protection

If an autonomous satellite processes personal data, GDPR issues may arise.

Consumer and commercial law

Satellite-navigation and communication services can create contractual and consumer claims.

5. Autonomous Spacecraft Does Not Mean Autonomous Legal Responsibility

This is the most important principle.

An autonomous spacecraft might independently decide:

"Change orbital trajectory."

But legally the question remains:

Who designed, supplied, operated, supervised or controlled the system?

Potentially responsible parties include:

spacecraft manufacturer;

satellite operator;

launch provider;

software developer;

AI developer;

ground-control operator;

maintenance contractor;

navigation-system provider;

cybersecurity provider;

owner;

launching State;

insurer.

Therefore:

Autonomous decision ≠ autonomous legal personality.

6. Elements of an Autonomous Space-System Liability Claim

A useful examination framework is:

A-F-C-D-R

A — Actor

Identify the potentially responsible party.

F — Failure

Identify the failure:

hardware failure;

software defect;

AI error;

sensor failure;

inadequate training;

defective update;

cybersecurity weakness;

communication failure;

poor maintenance.

C — Causation

Show that the system failure caused the harmful event.

D — Damage

Identify:

personal injury;

property damage;

satellite damage;

environmental damage;

economic loss;

service interruption;

consequential losses.

R — Remedy

Possible remedies include:

damages;

repair/replacement;

contractual compensation;

injunction;

insurance recovery;

contribution between responsible parties.

7. Autonomous Collision Liability

One of the most important scenarios is:

Autonomous satellite A automatically changes trajectory and collides with satellite B.

The legal analysis requires examination of:

whether the trajectory decision was reasonably foreseeable;

whether collision-avoidance software was defective;

whether sensors supplied accurate information;

whether the operator received warnings;

whether human intervention was reasonably possible;

whether another satellite contributed to the collision;

whether the collision resulted from a cyberattack;

whether the operator complied with applicable space-safety requirements.

A collision involving autonomous systems creates a multi-layer causation problem.

For example:

Sensor error → incorrect AI prediction → autonomous manoeuvre → collision → satellite destruction → commercial loss.

Each stage may involve a different legal actor.

8. Product Liability and Autonomous Space Hardware

European product-liability principles are especially important where the spacecraft or its components are defective.

The traditional EU Product Liability Directive established strict liability for defective products.

The newer EU Product Liability Directive 2024/2853 expands the technological framework and is particularly relevant to modern software and digital products.

This is significant because autonomous space systems may involve software-driven products rather than merely mechanical objects.

EU policy materials have specifically recognised that autonomous and interconnected systems create difficulties concerning:

definition of product;

definition of defect;

allocation of liability;

interconnected products and services;

autonomous decision-making. (Eur-Lex)

9. Case Law

Case 1 — Boston Scientific Medizintechnik GmbH v AOK Sachsen-Anhalt

Joined Cases C-503/13 and C-504/13

This is one of the most useful European authorities for autonomous-space-system liability by analogy.

The case concerned pacemakers and implantable cardioverter defibrillators. The Court considered products forming part of a production series where there was an increased risk of failure.

The Court accepted that products belonging to a group or series presenting an abnormally high risk of failure could be treated as defective without requiring proof that every individual product had actually malfunctioned. (Eur-Lex)

Relevance to autonomous spacecraft

Suppose an operator discovers that a particular model of autonomous satellite contains a systematic software defect capable of causing dangerous orbital manoeuvres.

The reasoning in Boston Scientific can support the proposition that systemic safety risks can be legally important even before every individual unit causes actual damage.

Principle

A systematic safety defect can be legally significant even when the particular unit has not yet produced catastrophic damage.

10. Case 2 — O'Byrne v Sanofi Pasteur

Case C-127/04

The CJEU considered when a product is regarded as having been put into circulation for purposes of product liability. The dispute involved supply by a producer to its wholly owned subsidiary. (Eur-Lex)

Relevance to space systems

This becomes important because a spacecraft can pass through multiple organisational stages:

Manufacturer → parent company → launch provider → operator → customer.

A claimant may therefore need to determine:

who manufactured the spacecraft;

who released it;

when control passed;

who operated it;

whether the defect existed before launch;

whether later software modifications created the defect.

Principle

The identification of the relevant producer and the moment of putting the product into circulation can be critical to liability.

11. Case 3 — Centre hospitalier universitaire de Besançon v Dutrueux

Case C-495/10

The CJEU considered the relationship between EU product liability and national liability rules where defective equipment was used by a service provider.

The Court held that the Product Liability Directive did not prevent a Member State from maintaining a national liability regime applicable to a service provider, provided the Directive's product-liability framework remained available against the producer where its conditions were fulfilled. (Eur-Lex)

Relevance to autonomous spacecraft

Space activities frequently combine products and services.

For example:

Autonomous satellite + ground-control service + cloud software + maintenance service.

A claimant should therefore not assume that the manufacturer's product-liability regime is the only possible route.

A satellite operator or service provider may potentially face liability under separate contractual or national civil-law rules.

Principle

Product liability and service-provider liability can coexist under appropriate national legal regimes.

12. Case 4 — Skov Æg v Bilka

Case C-402/03

The CJEU considered liability of suppliers under the Product Liability Directive. (Eur-Lex)

Relevance to autonomous space systems

A spacecraft is rarely produced by one entity.

A satellite may contain:

propulsion supplied by Company A;

sensors supplied by Company B;

AI software supplied by Company C;

communications hardware supplied by Company D;

integration performed by Company E.

If the autonomous spacecraft fails, determining whether a supplier can be treated as legally responsible becomes important.

Principle

Liability under the harmonised product-liability regime must be analysed according to the statutory categories of producer and supplier; contractual supply-chain arrangements do not automatically determine statutory liability.

13. Case 5 — Veedfald v Århus Amtskommune

Case C-203/99

This case concerned a defective product used in the context of medical treatment and the interpretation of the Product Liability Directive.

It is important because the Court adopted a functional approach to whether a product can fall within the product-liability regime even when it is used in providing a service.

Relevance to autonomous space operations

An autonomous satellite may be simultaneously:

a physical product;

a software-controlled machine;

an infrastructure component;

a service platform.

The fact that the spacecraft is being used to provide a service does not necessarily eliminate product-liability questions.

Principle

The distinction between product and service must be examined carefully rather than assumed from the commercial structure.

14. Case 6 — Galileo International Technology and Others v Commission

Case T-279/03

This is particularly valuable because it directly involved the Galileo satellite-navigation programme.

The applicants alleged that the European Community's use of the name "Galileo" caused damage to their trademark and commercial interests.

The General Court considered the requirements for non-contractual liability of the Community, including causation. It stressed that the alleged damage must be a sufficiently direct consequence of the conduct complained of. (Infocuria)

Relevance to autonomous space systems

The case provides an important European space-related liability authority for causation and attribution.

Suppose:

EU satellite programme → autonomous system → technical event → third-party commercial loss.

The claimant cannot simply show that the loss occurred during the operation of the space programme. The claimant must establish a legally sufficient causal connection between the defendant's conduct and the damage.

Principle

Mere connection with a space project is insufficient; legally relevant causation must be established.

15. Case 7 — Airbus Defence and Space SAS v European Defence Agency

Case T-105/24, General Court, judgment of 1 July 2026

This recent case concerned a public procurement procedure involving satellite communications, equipment and related services. The General Court considered issues including equal treatment, assessment of tenders, non-contractual liability and loss of opportunity. (Curia)

Relevance

Although it was not a spacecraft-accident case, it illustrates an important aspect of European space-system liability:

Space liability can arise not only from physical accidents but also from:

procurement decisions;

space-system contracts;

technical specifications;

tender evaluation;

loss of commercial opportunity.

This is particularly relevant to complex autonomous-space programmes involving governments, ESA, EU bodies and private aerospace companies.

16. Case 8 — Galileo/European Satellite Programme Liability

The broader Galileo litigation and EU regulatory framework is also important because Galileo and EGNOS involve management, operation, maintenance, continuous improvement, evolution and protection of space- and ground-based infrastructure. (Eur-Lex)

The European legal framework therefore treats space systems as complex interconnected infrastructures rather than isolated spacecraft.

For autonomous systems, this means liability may extend across:

Satellite → ground station → communications network → software → control centre → service provider.

17. Fault in an Autonomous Space System

For a fault-based claim, possible forms of fault include:

1. Design fault

The autonomous decision architecture was unsafe.

2. Software fault

The algorithm incorrectly calculated:

orbital trajectory;

collision probability;

fuel consumption;

docking distance.

3. Sensor fault

The spacecraft received inaccurate information.

4. Training-data fault

An AI system was trained on inadequate or inappropriate data.

5. Update fault

A software update introduced a dangerous defect.

6. Maintenance fault

The operator failed to maintain the system.

7. Cybersecurity fault

The operator failed to protect the system against foreseeable cyber threats.

8. Supervision fault

The operator permitted excessive autonomy without adequate human monitoring.

18. Autonomous Decision-Making and Human Oversight

A major legal question is:

How much human supervision is required?

The answer depends on the system and applicable legislation.

For a high-risk spacecraft, reasonable precautions might include:

human override;

emergency shutdown;

collision-warning mechanisms;

redundant sensors;

independent navigation systems;

safe-mode operation;

software testing;

cybersecurity controls;

audit logs;

incident recording.

If an operator deliberately designs a spacecraft to operate autonomously, it cannot necessarily avoid responsibility merely by arguing:

"The computer made the decision."

European policy discussions on AI liability have specifically considered the possibility that an operator should not escape liability merely because harm resulted from autonomous activity of an AI system. (Eur-Lex)

19. Cyberattack and Autonomous Space Systems

Cybersecurity creates another difficult issue.

Imagine:

Hacker → compromises satellite software → autonomous satellite changes orbit → collision.

Potential defendants might argue:

third-party criminal intervention;

unforeseeable cyberattack;

force majeure;

absence of negligence;

absence of causation.

The claimant may respond that the operator:

failed to patch software;

used inadequate encryption;

ignored known vulnerabilities;

failed to isolate critical systems;

failed to maintain authentication controls.

Thus, the question becomes whether the cyberattack was an external intervening cause or whether inadequate cybersecurity itself constituted a contributing fault.

20. Autonomous Space Debris

Autonomous systems create an especially difficult problem concerning space debris.

For example:

Autonomous satellite fails → breaks apart → thousands of fragments created → other spacecraft damaged.

Potential liability questions include:

Who operated the spacecraft?

Was the malfunction foreseeable?

Was the satellite properly maintained?

Were collision-avoidance mechanisms operational?

Was the satellite safely deorbited?

Did software failure cause the breakup?

Did another object contribute to the accident?

Was the operator's conduct attributable to the launching State?

Space debris therefore creates both international responsibility and possible private civil liability.

21. Contractual Liability

A satellite operator may have contracts with:

launch companies;

spacecraft manufacturers;

software developers;

insurance companies;

telecommunications providers;

satellite customers;

ground-station operators.

A contract may contain:

performance standards;

availability guarantees;

maintenance obligations;

cybersecurity obligations;

indemnities;

limitation-of-liability clauses;

force-majeure provisions;

insurance requirements.

For example:

An autonomous satellite is contractually required to maintain 99.9% service availability but repeatedly shuts down because of defective autonomous fault-management software.

The customer may have a contractual claim even if a separate product-liability claim is difficult.

22. Tort/Delict Liability

A tort/delict claim generally requires some combination of:

legal duty;

breach;

causation;

damage.

The precise test differs among European jurisdictions.

Possible duties include:

duty to operate safely;

duty to maintain equipment;

duty to warn;

duty to prevent foreseeable collisions;

cybersecurity duty;

professional-standard duty.

23. Defective Autonomous Software

A particularly difficult issue is whether autonomous software is itself a defective product.

Modern European product-liability developments increasingly recognise the importance of:

software;

digital products;

AI;

updates;

cybersecurity;

interconnected systems.

This is especially significant for autonomous spacecraft because the physical satellite may remain unchanged while its behaviour changes after a software update.

Example

Before update:

Satellite behaves safely.

After update:

Collision-avoidance algorithm incorrectly classifies another spacecraft as non-threatening.

If damage results, investigators may need to determine:

whether the update was defective;

who supplied it;

who authorised it;

whether testing was adequate;

whether the operator could have prevented installation;

whether the software supplier remained responsible after deployment.

24. Causation Problems

Autonomous systems create multi-causal accidents.

Example:

Poor sensor → AI misclassification → faulty trajectory → operator fails to intervene → collision → satellite destruction.

There may be four potential causes.

A court may need to allocate responsibility among:

The Galileo litigation is useful for the general principle that the damage must have a sufficiently direct causal relationship with the defendant's conduct. (Eur-Lex)

25. Damage Recoverable

Potential damages include:

Physical damage

destruction of spacecraft;

damage to launch infrastructure;

damage to ground equipment;

personal injury.

Economic damage

lost satellite capacity;

telecommunications interruption;

lost commercial contracts;

launch replacement costs;

repair costs.

Consequential losses

business interruption;

loss of customers;

delayed scientific missions.

However, pure economic loss is treated differently across European national legal systems and may be restricted in tort.

Therefore, contractual claims can sometimes be especially important for commercial satellite services.

26. Insurance

Space insurance is extremely important because autonomous-space accidents can involve enormous losses.

Insurance may cover:

launch failure;

in-orbit failure;

satellite damage;

third-party liability;

business interruption.

An insurer may later pursue subrogation against a manufacturer, operator or contractor responsible for the loss.

Thus, a spacecraft accident can produce litigation between:

Owner ↔ Insurer ↔ Manufacturer ↔ Operator ↔ Software provider ↔ Launch provider.

27. Multiple Autonomous Systems and Chain Liability

Future space systems will increasingly operate as interconnected networks.

Example:

Satellite A communicates with Satellite B → B sends instruction to C → C changes orbit → C collides with D.

This creates a question of chain causation.

A claimant may need to identify:

which autonomous decision was decisive;

whether each system behaved within expected parameters;

whether communication between systems was foreseeable;

whether interoperability was adequately tested;

whether one operator should have anticipated another system's behaviour.

This is similar to the European Commission's broader concern that interconnected autonomous systems can make the precise source of damage difficult to establish. (Eur-Lex)

28. Evidentiary Problems

Autonomous-space litigation will heavily depend on technical evidence.

Important evidence includes:

flight logs;

telemetry;

AI decision logs;

sensor records;

software versions;

update history;

ground-control records;

cybersecurity logs;

maintenance records;

design documents;

testing reports;

collision-avoidance calculations;

communications between operators;

insurance records.

Black-box problem

If an AI system cannot adequately explain why it made a particular orbital manoeuvre, proving negligence or defect may become difficult.

This makes logging and traceability extremely important.

29. Defences

A defendant may raise:

A. No defect

The spacecraft performed according to its specifications.

B. No negligence

All reasonable precautions were taken.

C. No causation

The damage resulted from another spacecraft or external event.

D. Third-party interference

A hacker or other third party caused the event.

E. Force majeure

The event was extraordinary and unavoidable.

F. Contributory fault

The claimant's own spacecraft or operator contributed to the collision.

G. Contractual limitation

A valid limitation or exclusion clause may restrict contractual recovery, subject to mandatory law.

H. Scientific/technical uncertainty

The state of technology may make a particular failure difficult to predict, although this does not automatically eliminate liability.

30. Manufacturer vs Operator Liability

The actual allocation depends on the applicable law and contractual structure.

31. International Liability vs Private Civil Liability

These should not be confused.

International space liability

Usually concerns:

State → State

under international space treaties.

Private civil liability

Usually concerns:

Person/company → company/person

under:

contract;

tort/delict;

product liability;

consumer law;

insurance law.

A single accident may generate both levels simultaneously.

32. Important Legal Principle from the Galileo Litigation

The Galileo case is particularly useful because it demonstrates that the fact that conduct occurs within a major European space programme does not automatically establish compensation.

A claimant still has to establish the legal requirements for liability, including the necessary causal connection between the relevant conduct and the damage. (Eur-Lex)

Therefore:

Space activity + damage ≠ automatic civil liability.

33. Special Problem: AI That Learns After Launch

An autonomous satellite may change its behaviour through:

machine learning;

software updates;

adaptive algorithms;

autonomous optimisation.

Suppose:

Manufacturer delivers a safe system → system learns from operational data → autonomous algorithm develops unsafe behaviour → collision occurs.

The court may have to determine:

whether the original system was defective;

whether the learning behaviour was foreseeable;

whether the operator had monitoring obligations;

whether the manufacturer had continuing update obligations;

whether the software became defective later;

who controlled the learning process.

This represents one of the hardest future liability problems in European space law.

34. Special Problem: Autonomous Space Mining

Future autonomous systems may perform:

asteroid prospecting;

lunar excavation;

resource extraction;

robotic construction.

Possible disputes include:

damage to another operator's equipment;

interference with another mission;

property damage;

contract disputes;

resource-extraction agreements;

environmental concerns;

cross-border jurisdiction.

The traditional concept of a spacecraft as a passive object therefore becomes increasingly inadequate as autonomous systems become active participants in space operations.

35. Practical Liability Matrix

36. Six Core Case-Law Principles

The first six satisfy the requested minimum; the Airbus case provides an additional contemporary space-sector authority. The Galileo case is especially distinctive because it directly concerned the European satellite-navigation programme. (Infocuria)

37. Exam-Ready Legal Test

For an Autonomous Space System Liability Claim, remember:

S-A-F-C-D-R

S — Space system
What autonomous spacecraft or component was involved?

A — Actor
Who manufactured, operated, controlled or supplied it?

F — Failure
What went wrong?

C — Causation
Did that failure legally cause the damage?

D — Damage
What physical, property or economic loss occurred?

R — Remedy
Which international, contractual, tortious, product-liability or insurance remedy applies?

38. Conclusion

European law presently does not create a general rule under which an autonomous spacecraft becomes independently liable merely because it made an autonomous decision.

Instead, liability is generally attached to legally recognised actors:

manufacturers;

operators;

owners;

software providers;

service providers;

maintenance contractors;

launch providers;

and, at the international level, launching States.

The most important legal difficulty is causation and attribution. An autonomous spacecraft may make the immediate decision, but the court must investigate the entire technological chain: design → software → sensors → training/update → operator configuration → autonomous decision → harmful event → damage.

The European cases such as Galileo International Technology, Boston Scientific, O'Byrne, Dutrueux, Skov Æg and Veedfald provide the principal analogical foundations. Direct European judicial precedent specifically dealing with an AI-controlled spacecraft collision remains limited, so these authorities should be described as analogical European case law rather than direct autonomous-spacecraft precedents. (Eur-Lex)

Short rule for examination

An autonomous space system does not automatically become a separate legal person or liability bearer. European liability is generally determined by identifying the responsible human or legal entity, the applicable international or national liability regime, the defect or wrongful conduct, causal connection and legally recoverable damage. Automation changes how the harmful event occurs; it does not by itself eliminate the underlying legal duties.