Civil Law And Evidence Expert Reports And Burden Of Proof In Uae .
1. Introduction
The relationship between civil law, autonomous artificial intelligence (AI) and machine moral responsibility is an emerging area of legal scholarship in the United Arab Emirates (UAE).
Autonomous intelligence systems are increasingly capable of performing tasks traditionally undertaken by human beings, including:
Making financial and commercial decisions.
Operating autonomous vehicles.
Diagnosing medical conditions.
Managing industrial machinery.
Executing contractual obligations.
Evaluating insurance claims.
Conducting automated recruitment.
Controlling smart buildings and infrastructure.
Making decisions through machine-learning algorithms.
These developments raise an important legal question:
When an autonomous intelligence system makes an independent decision that causes injury, financial loss, discrimination or contractual damage, who should bear civil liability and moral responsibility?
Traditional civil law generally recognises natural persons and legally recognised organisations as subjects of rights and obligations. An AI system, however sophisticated, does not automatically acquire legal personality or become independently responsible for its actions.
In the UAE, responsibility must therefore be examined through existing principles of civil liability, contractual obligations, negligence, causation, product responsibility and statutory regulation.
The subject has five major dimensions:
Legal responsibility for autonomous decisions.
Ethical responsibility for machine-generated conduct.
Civil liability of developers, manufacturers and operators.
Human accountability for algorithmic decisions.
Future recognition or rejection of independent AI legal personality.
2. Meaning of autonomous intelligence systems
An autonomous intelligence system is a technological system capable of processing information, evaluating alternatives and executing decisions with varying degrees of human supervision.
There are four principal categories.
| Category | Characteristics | Example |
|---|---|---|
| Assisted intelligence | Provides recommendations | AI medical diagnostic assistant |
| Semi-autonomous intelligence | Acts with human approval | Automated financial approval |
| Highly autonomous intelligence | Makes decisions with limited supervision | Autonomous industrial machinery |
| Networked autonomous intelligence | Coordinates with other systems | AI-controlled logistics networks |
The legal difficulty increases as human intervention decreases.
For example, where a driver directly controls a vehicle, responsibility can ordinarily be assessed through established driving and negligence principles.
However, where an autonomous vehicle independently identifies a pedestrian, calculates its route and controls braking, determining responsibility may require investigation of the manufacturer, software developer, maintenance provider, vehicle operator and other relevant actors.
3. Legal framework governing AI responsibility in the UAE
A. Federal Decree-Law No. 5 of 1985 on Civil Transactions
The UAE Civil Transactions Law provides the fundamental principles for civil liability.
Its general framework recognises responsibility for harmful conduct and establishes the importance of causation and legally compensable damage.
Article 282 establishes the general principle that any person who causes harm to another must make good the damage.
This provision is particularly important for autonomous intelligence systems.
If an AI-enabled machine causes injury, the court must investigate the legally responsible person or organisation behind the system.
Potentially responsible parties include:
The manufacturer.
The software developer.
The system integrator.
The operating company.
The maintenance provider.
The entity deploying the AI system.
Article 283 addresses direct and indirect causation.
This distinction is important where damage results from several interacting causes, such as defective software, inadequate maintenance and negligent supervision.
Article 292 concerns the assessment of compensation and provides the basis for examining the actual harm suffered.
Article 293 addresses compensation for moral damage, including harm to dignity, reputation and other protected personal interests.
Together, these provisions provide a foundation for dealing with damage caused through autonomous technologies.
B. Federal Decree-Law No. 35 of 2022 on Evidence in Civil and Commercial Transactions
AI liability disputes frequently involve complicated digital evidence.
This legislation is important for determining:
Whether algorithmic logs are admissible.
Whether electronic records are authentic.
Whether AI-generated reports can be relied upon.
Whether technical experts are necessary.
Whether electronic communications establish contractual conduct.
The court must distinguish between an AI system's output and reliable evidence establishing the facts behind that output.
C. Federal Decree-Law No. 46 of 2021 on Electronic Transactions and Trust Services
This legislation recognises electronic transactions and electronic signatures.
It is relevant where autonomous systems enter into or execute digital transactions.
For example, an automated purchasing system may generate orders, accept commercial terms or communicate contractual consent.
However, the legal recognition of electronic transactions does not automatically make the AI system a separate legal person.
The identity and authority of the person or organisation operating the system remain important.
D. Federal Decree-Law No. 45 of 2021 on the Protection of Personal Data
AI systems frequently process personal information.
The Personal Data Protection Law is relevant where autonomous systems:
Collect personal information.
Profile individuals.
Make automated assessments.
Process sensitive personal information.
Transfer personal data across borders.
Use information without an adequate legal basis.
The law's principles of lawful and fair processing, transparency, security and data-subject rights are important safeguards against unethical AI deployment.
E. UAE National Strategy for Artificial Intelligence 2031
The UAE's AI strategy establishes a policy framework for developing AI capabilities across government and economic sectors.
Its relevance to civil law lies in the growing need to balance technological innovation with:
Public safety.
Human welfare.
Accountability.
Responsible innovation.
Institutional transparency.
Ethical AI governance.
A national AI development strategy should not, however, be confused with a statute independently creating civil liability for AI systems.
4. Machine moral responsibility versus legal responsibility
One of the most important theoretical questions is whether an autonomous machine can itself be morally or legally responsible for its conduct.
These concepts must be separated.
A. Machine moral responsibility
Machine moral responsibility concerns whether an AI system can be considered morally accountable for decisions that cause harm.
For example:
An autonomous medical system recommends withholding a particular treatment from a patient. The recommendation is based on an algorithm that systematically undervalues certain categories of patients.
The ethical questions include:
Can the AI system understand the consequences of its recommendation?
Can it distinguish between right and wrong?
Can it understand human suffering?
Can it recognise discrimination?
Can it be morally blamed for an unjustified decision?
Current AI systems can process ethical rules and simulate moral reasoning. However, this does not establish that they possess consciousness, genuine moral understanding or the capacity to experience responsibility.
B. Legal responsibility
Legal responsibility is different.
A person or organisation can be legally responsible even if it did not personally perform every technical operation that caused the damage.
For example:
A manufacturer may be liable for a defective autonomous machine.
A developer may be responsible for negligent software design.
An operator may be responsible for inadequate supervision.
A company may be liable for harmful deployment of an AI system.
Under the existing UAE framework, an AI system does not automatically become an independent civil debtor merely because it makes an autonomous decision.
C. Comparison
| Question | Machine moral responsibility | Civil legal responsibility |
|---|---|---|
| Can AI make decisions? | Yes | Yes |
| Can AI cause damage? | Yes | Yes |
| Can AI be morally blamed? | Philosophically disputed | Not necessary for liability |
| Can AI pay compensation independently? | No established mechanism | No general AI-specific legal-personality regime |
| Can developers be responsible? | Ethical responsibility | Potential civil liability |
| Can operators be responsible? | Ethical responsibility | Potential civil liability |
The central distinction is that the machine may generate the harmful decision, but the law can attach responsibility to the humans and organisations that design, control, deploy or operate it.
5. Dubai's autonomous vehicle legislation: a direct statutory example
Particularly important UAE legislation
Dubai has adopted a specific regulatory framework for autonomous vehicles through Law No. 9 of 2023 Regulating the Operation of Autonomous Vehicles in the Emirate of Dubai.
This legislation is especially important because it directly addresses the civil liability consequences of autonomous machine operation.
Article 2: Definition of operator
The legislation recognises an operator as a natural or legal person authorised by the Roads and Transport Authority (RTA) to conduct the relevant autonomous-vehicle activity.
This is significant because responsibility is connected to an identifiable legal person rather than being assigned to the autonomous vehicle as an independent moral actor.
Article 11: Obligations of operators
Operators must:
Comply with licensing requirements.
Maintain autonomous vehicles.
Provide appropriate passenger instructions.
Ensure compliance with operational requirements.
Report accidents.
Respond to malfunctions.
Provide relevant information to competent authorities.
These obligations establish human and organisational accountability around autonomous technology.
Article 12: Obligations of agents
Authorised agents also have regulatory obligations concerning autonomous vehicles.
This recognises that responsibility can arise at different points in the technological supply chain.
Article 14: Civil liability
This is the most significant provision for the present topic.
Article 14(a) provides that the operator is liable to compensate for property damage or personal injury caused by the autonomous vehicle.
It also preserves the operator's right to seek recourse against the person actually responsible under the general rules of liability.
Article 14(b) excludes third-party liability of the RTA for damage resulting from the use of autonomous vehicles.
Legal significance:
The legislation establishes a particularly useful allocation of responsibility:
The autonomous vehicle performs the operation.
The operator bears the specified statutory compensation responsibility.
The operator may pursue the person actually responsible.
General civil liability principles determine the underlying responsibility.
This is a practical alternative to treating an autonomous vehicle as a separate legal person.
Dubai Land Department+1
D. Implementing Resolution No. 939 of 2025
Administrative Resolution No. 939 of 2025 provides the implementing bylaw for Dubai's autonomous vehicle legislation.
It introduces important operational requirements, including:
Technical and safety standards.
Comprehensive insurance.
Secure storage of operational data.
Recording of movement and maintenance information.
Reporting of malfunctions.
Software-update controls.
Restrictions on unauthorised modifications.
Procedures for restoring control when the system malfunctions.
These provisions are significant because they make it easier to investigate how an autonomous system operated before an accident.
They also strengthen the possibility of identifying whether the damage arose from software defects, maintenance failures or operational negligence.
Dubai Land Department
E. Executive Council Resolution No. 14 of 2026
A further development is Dubai Executive Council Resolution No. 14 of 2026, issued on 12 May 2026.
It approves activities related to autonomous vehicles and their operational phases.
Among other matters, it addresses authorised use, operational safety, passenger requirements and responsibility for unauthorised activation of autonomous driving systems.
This reinforces the principle that autonomous technological operations must remain within a regulated framework of human and institutional accountability.
Dubai Land Department
6. Ethical principles governing autonomous intelligence systems in the UAE
The UAE's AI Ethics Principles and Guidelines provide an important policy foundation.
The framework identifies several principles relevant to machine moral responsibility.
A. Human-centred AI
AI should be developed to serve human welfare rather than operate without regard to human interests.
B. Fairness and impartiality
AI systems used in important decisions should avoid unjustified discrimination.
C. Transparency
Users should have meaningful information about how AI systems are used, particularly when decisions affect their rights and interests.
D. Safety and security
AI systems should be designed to avoid foreseeable harm and should undergo appropriate safety evaluation.
E. Human control
Decisions involving life and death should not be delegated to AI systems without effective human control.
F. Accountability
Organisations involved in the design, development and deployment of AI should accept responsibility for the consequences of their systems.
These principles are especially important where AI systems operate in healthcare, transportation, banking and public administration.
The UAE's Charter for the Development and Use of Artificial Intelligence further emphasises responsible AI, privacy, security and equitable technological access.
Artificial Intelligence Office, UAE+1
7. Civil liability of developers, manufacturers and operators
The central practical issue is determining which participant in the AI ecosystem should compensate the victim.
A. Liability of AI developers
Developers may face civil liability where damage is connected with failures such as:
Negligent algorithmic design.
Inadequate testing.
Failure to address foreseeable risks.
Incorrect training data.
Failure to correct known software vulnerabilities.
Inadequate safeguards against discriminatory outputs.
For example, if a developer knows that an autonomous medical system regularly produces dangerous recommendations but continues distributing it without appropriate warnings or corrections, its conduct may become relevant to a civil liability claim.
B. Liability of manufacturers
Manufacturers may face liability where the physical product or integrated system is defective.
Potential defects include:
Faulty sensors.
Defective braking systems.
Inadequate safety mechanisms.
Incorrect hardware integration.
Failure to implement necessary software updates.
The relevant legal analysis must distinguish manufacturing defects, design defects and failures to warn.
C. Liability of operators
Operators may be responsible where they:
Ignore mandatory safety requirements.
Fail to maintain autonomous equipment.
Deploy systems outside authorised conditions.
Ignore known system malfunctions.
Fail to provide necessary human supervision.
Dubai Law No. 9 of 2023 provides a specific statutory example of operator responsibility for autonomous vehicles.
D. Liability of corporate deployers
A company deploying AI in its business may face liability where its own decisions contribute to harm.
Examples include:
Using an inadequately tested AI recruitment system.
Deploying an unreliable financial assessment model.
Using an unsafe autonomous industrial system.
Relying on automated medical decisions without appropriate clinical safeguards.
The legal assessment depends on the applicable statutory duties, contractual obligations, causation and evidence.
8. At least seven important judicial decisions
There is an important research limitation: UAE reported case law has not yet established a general doctrine under which an autonomous AI system is personally morally or civilly responsible for its decisions.
The following cases therefore combine UAE civil-liability authorities with comparative decisions that help develop the legal principles. They should not be misrepresented as UAE judgments specifically concerning AI.
Case 1. Abu Dhabi Court of Cassation — Civil Appeal No. 96/2024
UAE mainland authority
Subject: General civil liability under Article 282
This judgment is identified among the decisions applying Article 282 of the UAE Civil Transactions Law.
Legal relevance:
Article 282 establishes the general principle of compensation for harm caused to another.
In autonomous intelligence disputes, this principle provides the foundation for investigating the legally responsible person or organisation behind harmful machine operations.
Application:
If an AI-controlled industrial machine injures an employee, the court can examine the conduct of the manufacturer, maintenance contractor and operating company.
The case should be used as a general civil-liability authority, not as an AI-specific precedent.
Case 2. Abu Dhabi Court of Cassation — Commercial Appeal No. 1137/2023
UAE mainland authority
Subject: Harm and assessment of compensation
This decision is identified in connection with Articles 282 and 292 of the Civil Transactions Law.
Legal relevance:
The civil-liability framework requires examination of the damage suffered and the appropriate compensation.
Application:
Where an autonomous financial system executes an incorrect transaction, the court must examine:
Actual financial loss.
Causation.
Lost profits where legally recoverable.
The conduct of responsible parties.
The case provides a statutory-liability reference for analysing financial damage associated with automated systems.
Case 3. Dubai Court of Cassation — Commercial Appeal No. 151/2025
UAE mainland authority
Subject: Compensation for harmful conduct
This decision is identified in connection with Articles 282 and 292 of the Civil Transactions Law.
Legal relevance:
It is relevant to the relationship between harmful conduct and compensation.
Application:
Where an autonomous commercial agent executes an unauthorised transaction and causes financial damage, the court must determine the applicable liability and the extent of legally recoverable loss.
The machine's autonomy does not automatically eliminate the responsibility of the relevant legal persons.
Case 4. Donoghue v Stevenson [1932] AC 562
Comparative authority — United Kingdom
Subject: Manufacturer's duty of care
This landmark decision established the modern neighbour principle in negligence.
The case concerned a consumer who suffered illness after consuming ginger beer containing a decomposed snail.
The House of Lords recognised a duty of care owed by manufacturers to ultimate consumers in appropriate circumstances.
Legal principles:
Manufacturers may owe duties to consumers.
Foreseeable harm can establish a duty of care.
A manufacturer cannot necessarily escape responsibility merely because there is no direct contractual relationship with the injured consumer.
Application to autonomous intelligence:
An AI developer or manufacturer may owe duties to foreseeable users and third parties exposed to risks created by its product.
For example, a company manufacturing an autonomous delivery robot must consider foreseeable risks to pedestrians.
Case 5. Grant v Australian Knitting Mills Ltd [1936] AC 85
Comparative authority — Privy Council
Subject: Defective products and consumer protection
The case involved a consumer who suffered dermatitis after wearing woollen underwear containing excessive chemical residue.
The Privy Council upheld the manufacturer's responsibility in the circumstances.
Legal principles:
Product defects can give rise to manufacturer liability.
Consumers may recover for injury caused by defective products.
A manufacturer must take appropriate care in producing goods for consumer use.
Application to autonomous intelligence:
This reasoning is relevant where an AI-enabled product causes harm because of an unsafe design or manufacturing defect.
For example, a defective autonomous medical device may cause injury because its integrated control system fails to operate safely.
Case 6. Bolam v Friern Hospital Management Committee [1957] 1 WLR 582
Comparative authority — United Kingdom
Subject: Professional negligence and technical standards
The case concerned alleged medical negligence and the standard of care expected from medical professionals.
The decision established the traditional Bolam approach to professional standards.
Legal principles:
Professional conduct may be assessed against accepted professional practice.
Technical expertise is relevant when determining whether a professional acted negligently.
Professional standards do not automatically eliminate judicial examination of the conduct in question.
Application to autonomous intelligence:
Consider a hospital using AI-assisted diagnosis.
If a doctor relies on an AI recommendation that contradicts obvious clinical indicators, the court may need to examine the professional standard of care.
The relevant question is not simply whether the AI system made an error, but whether the responsible professionals and institution acted reasonably.
Case 7. Montgomery v Lanarkshire Health Board [2015] UKSC 11
Comparative authority — UK Supreme Court
Subject: Informed consent and disclosure of material risks
The Supreme Court considered the obligation of medical professionals to inform patients about material risks and reasonable alternatives.
Legal principles:
Patients have an important right to make informed decisions.
Material risks must be appropriately disclosed.
Professional responsibility cannot be reduced to a purely technical medical assessment.
Application to autonomous intelligence:
A healthcare provider using an AI diagnostic or treatment system may need to consider whether patients should be informed of the system's role and material risks.
The decision is particularly relevant to human oversight and patient autonomy.
Case 8. Prygodicz v Commonwealth of Australia (No 2) [2021] FCA 634
Comparative authority — Australia
Subject: Automated decision-making and civil harm
This proceeding concerned Australia's automated Robodebt debt-collection system.
The system used income averaging to identify alleged social-security overpayments.
The litigation involved claims concerning unjust enrichment and negligence.
The Federal Court approved a class-action settlement.
Legal significance:
Automated calculations can produce legally invalid outcomes.
Government organisations cannot assume that computer-generated assessments are legally correct.
Human and institutional accountability remains important.
Automated decision-making must comply with the legal rules governing the underlying decisions.
Application to the UAE:
A UAE organisation using AI to determine contractual debts, insurance claims or financial obligations cannot treat an automated output as a substitute for legal verification.
Important qualification: The case involved a settlement and approval of that settlement. It should not be cited as a final judicial determination establishing negligence liability for AI systems.
UAE Legislation+2
9. Comparative legal analysis of the cases
| Case | Main principle | AI-related application |
|---|---|---|
| Abu Dhabi Cassation 96/2024 | General tort liability | Responsibility for harmful AI deployment |
| Abu Dhabi Cassation 1137/2023 | Damage and compensation | AI-generated financial loss |
| Dubai Cassation 151/2025 | Compensation for harm | Autonomous commercial transactions |
| Donoghue v Stevenson | Manufacturer's duty | AI product safety |
| Grant v Australian Knitting Mills | Defective products | Defective autonomous devices |
| Bolam v Friern | Professional standard of care | AI-assisted medical negligence |
| Montgomery v Lanarkshire | Informed consent | Disclosure of AI involvement |
| Prygodicz v Commonwealth | Automated decision accountability | Unlawful algorithmic decisions |
These authorities collectively support the proposition that technological autonomy does not necessarily remove the legal duties of manufacturers, professionals, organisations or operators.
10. The philosophical foundations of machine moral responsibility
The concept of machine moral responsibility can be examined through four major philosophical approaches.
A. Deontological ethics
Deontological ethics focuses on duties, obligations and moral rules.
Applied to autonomous intelligence, it asks whether AI systems should be programmed to respect fundamental human rights and ethical restrictions.
For example:
An autonomous security robot should not unlawfully attack an innocent person simply because its algorithm identifies that person as a potential threat.
In the UAE, this approach supports the development of AI systems that respect human dignity, safety and applicable law.
B. Utilitarian ethics
Utilitarianism evaluates conduct according to its consequences for overall welfare.
For example, an autonomous vehicle may have to select between alternative emergency responses.
The system might calculate which response minimises expected injuries.
However, purely mathematical optimisation may create ethical difficulties if it disregards individual rights or treats particular individuals as less valuable.
UAE AI governance should therefore combine safety optimisation with legal protections for individuals.
C. Virtue ethics
Virtue ethics emphasises qualities such as fairness, responsibility, honesty and practical wisdom.
AI systems can be designed to follow ethical guidelines, but their ability to simulate virtuous behaviour does not establish that they possess human moral character.
The responsibility for designing and supervising ethical systems remains with human institutions.
D. Human-centred ethics
Human-centred ethics places human dignity, welfare and autonomy at the centre of technological development.
This approach is particularly relevant to UAE AI governance because it supports the principle that technology should improve human welfare rather than displace essential human accountability.
11. Hypothetical civil liability situations in the UAE
The following examples demonstrate how the principles could operate.
Scenario 1: Autonomous vehicle accident
An autonomous vehicle fails to identify a pedestrian and causes serious injury.
Potential liability: The operator under Dubai Law No. 9 of 2023, with possible recourse against the manufacturer, software developer or other responsible person.
От туризма до медицины: Сбер внедряет AI в ключевые отрасли Ростовской области
Scenario 2: AI medical misdiagnosis
An AI diagnostic system incorrectly identifies a serious disease, resulting in delayed treatment.
Potential liability: The healthcare provider, medical professional, developer or manufacturer, depending on the evidence and applicable duties.
Kontakt Reparatur Elektronisches Equipment | ERS Elektronik GmbH
Scenario 3: Autonomous industrial robot
An industrial robot incorrectly calculates the distance between a worker and a moving mechanical arm.
Potential liability: The employer, operator, manufacturer or maintenance provider, depending on causation and responsibility.
Institutional Algorithmic Trading Systems | Titan Quant
Scenario 4: Autonomous financial trading
An AI trading system executes unauthorised transactions and causes substantial financial losses.
Potential liability: The financial institution, authorised operator or other responsible party, depending on contractual authority and negligence.
The Rise of AI in Recruitment: What HR Leaders Must Know in 2025 | HR HUB
Scenario 5: Algorithmic employment discrimination
An automated recruitment system systematically rejects qualified applicants because of discriminatory training data.
Potential liability: The employer and potentially other responsible participants, subject to the applicable employment, data protection and civil liability rules.
12. The problem of identifying the responsible party
Autonomous intelligence systems frequently operate through interconnected technological networks.
A single harmful decision may involve several participants.
For example:
Autonomous AI responsibility network
AI developer
Algorithm design and training
Manufacturer / integrator
Hardware and software integration
Operator / deploying organisation
Operational supervision and safety
Autonomous system causes damage
Physical, financial or personal harm
Civil liability determination
Identify responsible parties and compensation
The main legal challenge is that no single participant may have complete control over the entire system.
A software developer may control the algorithm but not its deployment. A manufacturer may control hardware integration but not subsequent updates. An operator may control the environment but not the internal decision-making process.
UAE civil courts must therefore investigate the actual contribution of each party rather than automatically attributing all responsibility to one participant.
13. Defences and limitations of liability
Potential defences in autonomous intelligence litigation include:
1. Force majeure: Article 287 of the Civil Transactions Law recognises circumstances in which damage results from an external cause beyond the defendant's control, subject to the applicable legal requirements.
2. Victim contribution: The conduct of the injured party may be relevant to determining liability and compensation.
3. Third-party intervention: Damage may arise from an independent act of another person.
4. Proper maintenance: An operator may establish that it complied with relevant maintenance obligations, although this does not automatically eliminate other grounds of liability.
5. Absence of causation: A defendant may demonstrate that its conduct did not legally cause the damage.
6. Contractual allocation: Parties may allocate certain risks contractually, subject to mandatory law and applicable restrictions.
Importantly, the mere fact that an AI system behaved unexpectedly does not automatically establish force majeure.
An autonomous system's unpredictable behaviour may itself be a foreseeable technological risk requiring appropriate precautions.
14. Challenges for UAE civil courts
Several major challenges are likely to influence future litigation.
| Challenge | Legal difficulty |
|---|---|
| Algorithmic opacity | Understanding how a decision was generated |
| Machine-learning unpredictability | Identifying the source of unexpected conduct |
| Distributed responsibility | Identifying responsible participants |
| AI-generated evidence | Establishing authenticity and reliability |
| Moral responsibility | Distinguishing machine behaviour from human accountability |
| Cross-border development | Determining applicable law and jurisdiction |
| Software updates | Identifying which version caused the damage |
| Insurance coverage | Allocating losses arising from autonomous operations |
| Cyberattacks | Distinguishing system defects from external interference |
| Legal personality | Determining whether future AI entities should receive separate legal status |
15. Proposed reforms for UAE civil law
The following reforms could strengthen the UAE's approach to autonomous intelligence systems.
A. AI-specific civil liability framework
The UAE could consider legislation clarifying the responsibility of developers, manufacturers, integrators and operators.
Such legislation could establish specific rules for high-risk autonomous systems.
B. Mandatory AI liability insurance
High-risk autonomous systems could be subject to appropriate compulsory insurance requirements.
Dubai's autonomous vehicle implementing framework already provides an example through its comprehensive insurance requirement.
C. Mandatory human oversight
Systems affecting life, health, liberty and fundamental rights should have appropriate human oversight and intervention mechanisms.
D. Algorithmic audit obligations
High-risk AI systems could undergo independent technical audits covering:
Accuracy.
Safety.
Discrimination.
Data quality.
Security.
Reliability.
Explainability.
E. AI incident reporting
Organisations could be required to report serious AI-related incidents to competent authorities.
F. Preservation of algorithmic evidence
Developers and operators should preserve relevant technical records, including:
Decision logs.
Software versions.
Training and validation records.
System updates.
Operational data.
Human intervention records.
G. Clarification of AI legal personality
Any future consideration of AI legal personality should distinguish between technological autonomy and genuine legal accountability.
Granting legal personality to an AI system without assets, meaningful independent responsibility or an effective compensation mechanism could undermine victim protection.
For the foreseeable future, allocating liability to identifiable humans and organisations offers a more practical framework.
16. Final conclusion
Civil law and machine moral responsibility represent an important emerging legal challenge for the UAE.
Autonomous intelligence systems can make decisions, execute transactions and cause damage without direct human intervention at every stage. Nevertheless, the ability to make autonomous decisions does not establish that a machine possesses moral consciousness or independent legal responsibility.
The UAE's existing civil liability framework, particularly Articles 282, 283, 287, 292 and 293 of the Civil Transactions Law, provides a foundation for addressing harm caused by autonomous systems.
Dubai Law No. 9 of 2023 is especially important because Article 14 expressly allocates compensation responsibility to the operator of an autonomous vehicle while preserving recourse against the person actually responsible.
The UAE's AI ethical framework further supports responsible development, human welfare, safety and accountability.
The fundamental principle is:
Autonomous intelligence may perform an action independently, but technological independence must not become a means of escaping human and institutional legal accountability.
The future development of UAE civil law should therefore concentrate on transparent responsibility, effective human oversight, reliable technical evidence and meaningful compensation for victims of autonomous-system failures.
17. Important legislation and case-law revision
| Legal authority | Importance |
|---|---|
| UAE Civil Transactions Law, Article 282 | General liability for harm |
| Article 283 | Direct and consequential harm |
| Article 287 | External causes and defences |
| Article 292 | Assessment of compensation |
| Article 293 | Moral damage |
| Federal Decree-Law 35/2022 | Civil and commercial evidence |
| Federal Decree-Law 46/2021 | Electronic transactions |
| Federal Decree-Law 45/2021 | Personal data protection |
| Dubai Law 9/2023, Article 14 | Autonomous vehicle operator liability |
| Dubai Administrative Resolution 939/2025 | Autonomous vehicle implementation |
| Executive Council Resolution 14/2026 | Autonomous vehicle activities and operational phases |
Research note: The UAE Cassation decisions identified above are general civil-liability authorities, not reported judgments recognising machine moral responsibility. The comparative cases illustrate related principles rather than binding UAE AI law. This distinction is essential when using the material for academic research, legal writing or examination preparation.

comments