Civil Law And Ethics Of Autonomous Intelligence Systems And Machine Moral Accountability In Uae .

1. Introduction

The relationship between civil law, autonomous intelligence systems and machine moral accountability is an emerging area of UAE jurisprudence. It concerns the legal and ethical consequences of decisions made by artificial intelligence (AI) systems that operate independently or with limited human supervision.

Autonomous intelligence systems include:

Self-learning artificial intelligence and machine-learning systems.

Autonomous vehicles and intelligent transportation systems.

AI-powered medical diagnostic and treatment systems.

Automated financial and banking decision-making platforms.

Robotic systems used in construction and industrial operations.

AI-powered contractual negotiation and execution systems.

Autonomous agents capable of making economic and administrative decisions.

The central legal question is:

Who should bear civil responsibility when an autonomous machine makes a harmful, discriminatory, negligent or ethically unacceptable decision?

Under the UAE legal framework, artificial intelligence does not automatically acquire independent moral or civil legal responsibility merely because it can make decisions without continuous human intervention.

Responsibility generally remains connected to identifiable legal persons, including manufacturers, developers, operators, service providers, employers, owners and contracting parties.

The subject therefore requires examining three interconnected dimensions:

Civil liability for damage caused by autonomous intelligence.

Ethical accountability for machine-generated decisions.

Human and institutional responsibility for designing, deploying and supervising AI systems.

2. Legal framework governing autonomous intelligence in the UAE

A. UAE Civil Transactions Law

The UAE Civil Transactions Law is the foundation for determining civil liability arising from harmful conduct.

An important legislative development must be considered: Federal Decree-Law No. 25 of 2025 promulgating the new Civil Transactions Law. Its provisions should be distinguished from the earlier Federal Law No. 5 of 1985, which is the source of many traditionally cited civil-liability principles.

The relevant principles include:

1. Liability for harmful acts

The earlier Civil Transactions Law, Article 282, established that anyone causing harm to another must compensate the injured party, even where the actor lacks discernment.

The corresponding new-law provision is Article 246.

This principle is significant because an autonomous machine may cause physical or financial damage without conventional human decision-making at the precise moment of injury.

However, it does not automatically make the machine a legal person. Liability must be attributed under the applicable statutory framework to an identifiable natural or legal person.

2. Direct and indirect causation

The earlier Articles 283 and 284 distinguish direct harm from harm caused indirectly.

This distinction is important where:

A developer creates a defective algorithm.

A manufacturer incorporates the algorithm into a machine.

An operator deploys the machine.

The machine independently performs a harmful action.

The court must examine the causal relationship between each party's conduct and the resulting damage.

3. Deception and misleading conduct

The earlier Article 285 addresses harm resulting from deception.

This may be relevant where an AI provider falsely represents that its system is completely safe, unbiased, independently verified or capable of making accurate medical or financial decisions.

4. External causes and force majeure

The earlier Article 287 recognizes certain external causes, including force majeure, unavoidable accidents and acts of third parties, as potential grounds for excluding liability.

A developer cannot automatically escape liability by describing an algorithmic failure as an unexpected technological event. The court must examine whether the event was genuinely unavoidable and whether reasonable precautions were taken.

5. Compensation

Civil compensation may cover proven material loss and other legally recoverable damage, subject to the applicable statutory requirements and rules of causation.

B. Dubai autonomous vehicle legislation

A particularly important UAE legislative development is Dubai Law No. 9 of 2023 Regulating the Operation of Autonomous Vehicles.

Article 14 expressly provides that the operator is liable for compensating damage to property or harm to individuals caused by an autonomous vehicle.

The operator retains a right of recourse against the person actually responsible under the general rules of liability.

This is especially significant because it provides a statutory allocation of responsibility without requiring the injured person to establish that the vehicle itself possesses moral or legal personality.

C. UAE AI Ethics Principles and Guidelines

The UAE's AI Ethics Principles and Guidelines (December 2022) establish an ethical governance framework based on:

Fairness.

Accountability.

Transparency.

Explainability.

Resilience and reliability.

Safety and security.

Human-centred values.

Sustainability and privacy protection.

The UAE Charter for the Development and Use of Artificial Intelligence reinforces these objectives through human oversight, responsible governance, safety, privacy and transparency.

These principles are important when determining whether an organisation has adopted appropriate safeguards. However, ethical principles should not automatically be treated as independent causes of action or substitutes for statutory liability.

3. Meaning of machine moral accountability

Machine moral accountability refers to the question of whether an artificial intelligence system can be regarded as morally responsible for its own decisions.

Traditional moral accountability generally requires several elements:

Awareness of the consequences of an action.

Capacity to distinguish acceptable from unacceptable conduct.

Ability to understand moral obligations.

Freedom or meaningful control over decisions.

Capacity to accept responsibility for wrongdoing.

Advanced AI systems can simulate ethical reasoning, identify competing interests and generate decisions based on programmed values.

Nevertheless, simulated moral reasoning is not equivalent to legally recognised moral agency.

For example, an autonomous medical system might recommend denying treatment to a patient because its predictive model incorrectly classifies the patient's survival probability.

The machine may have generated the recommendation independently, but the relevant legal questions concern:

Who developed the predictive model?

Who selected its training data?

Who approved its clinical deployment?

Who was responsible for monitoring its accuracy?

Who had authority to override the recommendation?

Was the patient adequately informed?

Thus, machine moral accountability must be distinguished from human and institutional legal accountability.

4. Major ethical principles governing autonomous intelligence

A. Human dignity

AI systems must respect human dignity and individual rights.

An autonomous system should not be permitted to make decisions that arbitrarily undermine personal autonomy, bodily integrity or fundamental interests.

Examples include discriminatory recruitment, denial of essential services and unjustified restrictions on personal freedom.

B. Algorithmic fairness

Autonomous intelligence systems may reproduce discrimination contained in their training data.

For example, an AI recruitment platform may systematically disadvantage applicants because its training data reflects historical employment discrimination.

Civil liability may arise where a legally actionable wrong, damage and the necessary causal connection can be established.

C. Transparency and explainability

Organisations deploying AI should maintain sufficient information to explain consequential decisions.

Relevant documentation includes:

Algorithmic design and operational specifications.

Training-data sources and limitations.

Model updates.

Risk assessments.

Human intervention records.

Decision logs.

Incident reports.

A lack of explainability can create serious evidentiary difficulties during civil litigation.

D. Human supervision

Human supervision is especially important in high-risk sectors.

For example, autonomous systems used in healthcare, aviation, banking and industrial operations should have appropriate escalation and intervention mechanisms.

Human supervision must be meaningful rather than merely symbolic.

E. Prevention of foreseeable harm

Developers and operators should identify foreseeable risks before deployment.

This includes testing for algorithmic bias, cybersecurity vulnerabilities, system failures, unsafe updates and unintended interactions between autonomous systems.

5. Allocation of civil liability among AI stakeholders

The UAE framework requires examining the role of each participant in the AI lifecycle.

Responsible partyPotential basis of liability
AI developerDefective design, inadequate testing or foreseeable algorithmic risks
ManufacturerUnsafe hardware, integration defects or product-related failures
AI operatorImproper deployment, inadequate supervision or failure to intervene
EmployerLegally attributable acts or omissions of employees and agents
Data providerMisleading or unlawfully supplied data, where the relevant legal elements are established
AI service providerBreach of contractual obligations, inadequate security or negligent services
System ownerFailure to maintain, inspect or safely operate the system

A. Developer responsibility

Developers may face civil claims where defective programming, inadequate testing or foreseeable design deficiencies cause injury or financial loss.

For example, a developer creates an autonomous industrial robot that incorrectly identifies a human worker as an obstruction and performs a dangerous movement.

The court may investigate whether:

Appropriate safety testing was conducted.

Emergency shutdown mechanisms were incorporated.

The training data was sufficiently reliable.

Known software vulnerabilities were addressed.

The developer complied with applicable technical standards.

B. Operator responsibility

An operator may be responsible for deploying an AI system without adequate monitoring or safety controls.

Consider an autonomous delivery vehicle operating in Dubai.

If the vehicle causes an accident, Dubai Law No. 9 of 2023 provides an express liability framework under Article 14.

The operator may compensate the injured person and subsequently seek recovery from the person actually responsible.

C. Manufacturer responsibility

Manufacturers may be liable where defective hardware or unsafe integration contributes to the damage.

For example, an autonomous medical robot may malfunction because of a defective sensor.

The investigation should distinguish between:

Manufacturing defects.

Software defects.

Maintenance failures.

Improper installation.

External interference.

D. Employer responsibility

Where an employee improperly supervises or deploys an autonomous system, the employer's potential responsibility must be considered under the applicable rules governing employment, agency and attribution of harmful acts.

An organisation cannot necessarily avoid responsibility by arguing that its employee relied upon an automated recommendation.

E. Contractual responsibility

AI service agreements should clearly allocate responsibility for:

System reliability.

Data accuracy.

Cybersecurity.

Software updates.

Human intervention.

System downtime.

Error reporting.

Indemnification.

Insurance.

Compensation for legally recoverable losses.

Contractual provisions cannot automatically eliminate mandatory legal responsibilities or the rights of third parties.

6. Case laws and judicial authorities

Important qualification: There is no established body of reported UAE mainland judgments recognising autonomous AI systems as independent moral or civil legal persons. The following authorities are relevant because they establish principles concerning negligence, causation, technological transactions, professional responsibility, automated decision-making and legal accountability.

The DIFC decisions are jurisdiction-specific, while the foreign decisions are comparative authorities rather than binding UAE precedents.

Case 1. Gate Mena DMCC v Tabarak Investment Capital Limited [2020] DIFC TCD 001

DIFC Technology and Construction Division — 2022 judgment

This case concerned a cryptocurrency transaction involving the transfer of approximately 300 Bitcoin and the alleged failure of arrangements intended to safeguard the digital assets.

The claimants alleged negligence and breach of duties concerning the handling of the digital assets.

The court examined:

The existence of a duty of care.

Reasonable professional standards.

Foreseeability of financial loss.

Proximity between the parties.

Whether the defendants had assumed responsibility.

The negligence claims were dismissed.

Relevance to autonomous intelligence: An organisation participating in a technologically sophisticated transaction is not automatically liable merely because a loss occurs. The court must examine the actual obligations, conduct and causal connection.

Case 2. Gate Mena DMCC v Tabarak Investment Capital Limited [2023] DIFC CA 002

DIFC Court of Appeal — 2023

The appeal concerned the legal consequences of the cryptocurrency transaction and the duties allegedly owed by the defendants.

The Court of Appeal examined the requirements for establishing a duty of care, including:

Reasonable foreseeability.

Sufficient proximity.

Whether imposing a duty would be fair, just and reasonable.

The judgment is important because it demonstrates that technological complexity does not remove the ordinary legal requirements for negligence.

Relevance to autonomous intelligence: If an AI service provider gives assurances about the safety or reliability of an autonomous system, its responsibility may depend on whether it assumed a legally recognised duty and whether the claimant reasonably relied upon that responsibility.

Case 3. Gate Mena DMCC v Tabarak Investment Capital Limited [2024] DIFC DEC 002

DIFC Digital Economy Court — 17 June 2026 retrial judgment

This proceeding concerned the retrial of an issue arising from the earlier cryptocurrency dispute.

The court considered the contractual arrangements governing the custody and handling of Bitcoin and the extent of the defendant's obligations.

The claim was dismissed.

The decision illustrates the importance of determining the actual contractual allocation of responsibility rather than assuming that a participant guarantees every technological outcome.

Relevance to autonomous intelligence: AI deployment contracts should expressly identify which party is responsible for supervision, security, operational control and foreseeable system failures.

Case 4. Abu Dhabi Commercial Bank PJSC v Shetty & Others [2022] EWHC 529 (Comm)

English High Court — UAE law issues

This case involved proceedings in England in which the claimant relied on UAE law principles concerning harmful acts under Federal Law No. 5 of 1985.

The proceedings considered the pleaded application of Articles 282 and following of the UAE Civil Code.

The case is particularly useful for understanding how UAE civil liability principles may be presented and examined in cross-border commercial litigation.

Relevance to autonomous intelligence: Where an AI-generated action causes financial loss, the claimant must identify the relevant harmful conduct and establish the legally required relationship between that conduct and the damage.

Case 5. R (Bridges) v Chief Constable of South Wales Police [2020] EWCA Civ 1058

England and Wales Court of Appeal — 2020

This case concerned the police use of automated facial recognition technology in public spaces.

The court examined the legality of the deployment of facial recognition systems, including questions involving discretion, safeguards, data protection and equality.

The decision identified deficiencies in the framework governing the deployment of the technology.

Relevance to autonomous intelligence: Algorithmic systems can produce decisions that affect personal rights even when their operators do not intend discrimination.

The case supports the importance of:

Algorithmic fairness.

Appropriate human supervision.

Clear deployment policies.

Effective safeguards.

Accountability for automated decision-making.

Case 6. State v. Loomis, 881 N.W.2d 749 (Wis. 2016)

Wisconsin Supreme Court — 2016

This case involved the use of the COMPAS risk-assessment system in criminal sentencing.

The defendant challenged the use of an algorithmic assessment in the sentencing process.

The court permitted the consideration of the risk assessment subject to important limitations and warnings.

The case raised significant concerns regarding:

Transparency.

Proprietary algorithms.

Potential algorithmic bias.

Reliability of automated predictions.

Human responsibility for decisions influenced by machine-generated assessments.

Relevance to autonomous intelligence: An algorithm should not become an unquestionable decision-maker. Human authorities must critically evaluate the information produced by automated systems.

Case 7. Uber BV v Aslam [2021] UKSC 5

UK Supreme Court — 2021

This case concerned the employment status of Uber drivers and the extent of Uber's control over their working arrangements.

The court examined the practical operation of Uber's platform, including its control over fares, contractual arrangements and working conditions.

Relevance to autonomous intelligence: Algorithmic management does not necessarily remove the legal responsibility of the organisation controlling the platform.

Businesses cannot automatically avoid legal obligations by claiming that operational decisions are generated by software.

Case 8. Thaler v Comptroller-General of Patents, Designs and Trade Marks [2023] UKSC 49

UK Supreme Court — 2023

This case concerned whether an artificial intelligence system known as DABUS could be recognised as an inventor under patent legislation.

The Supreme Court upheld the requirement that an inventor under the relevant legislation must be a natural person.

Although this was an intellectual property dispute rather than a civil damages case, it is significant for the broader question of AI legal personality.

Relevance to autonomous intelligence: The ability of a machine to generate inventions or sophisticated decisions does not automatically establish that it possesses independent legal personality.

Overall judicial significance

These eight authorities support an important distinction:

Technological autonomy does not automatically create legal personality.

A machine-generated decision does not eliminate human or corporate responsibility.

Liability depends on applicable legal duties, causation, contractual obligations and statutory allocation.

Ethical concerns can inform governance and the assessment of reasonable precautions, but they do not independently establish civil liability in every case.

7. Application in important UAE industries

A. Healthcare and autonomous medical systems

AI-powered medical systems can diagnose illnesses, recommend treatments and assist in surgical procedures.

Potential civil disputes include:

Incorrect diagnosis.

Defective treatment recommendations.

Failure to identify medical emergencies.

Unsafe autonomous surgical operations.

Failure to disclose the use of AI.

A hospital may face liability where the deployment, supervision or clinical use of an AI system falls below applicable professional and legal standards.

AI should support, rather than automatically replace, legally required professional judgment.

B. Banking and financial services

Autonomous financial systems can determine creditworthiness, detect fraud and execute transactions.

Potential disputes include:

Unjustified loan rejection.

Discriminatory credit scoring.

Incorrect fraud detection.

Unauthorised automated transactions.

Algorithmic investment losses.

Banks and financial institutions must consider contractual obligations, applicable financial regulations, data protection and civil liability.

C. Autonomous transportation

Dubai Law No. 9 of 2023 is particularly important in this sector.

Its Article 14 provides a specific compensation mechanism against autonomous vehicle operators.

The implementing Administrative Resolution No. 939 of 2025 further addresses operational responsibilities, including malfunction notification, safety compliance and intervention when control is compromised.

D. Construction and industrial robotics

Autonomous machines used in construction may cause:

Structural damage.

Workplace injuries.

Equipment failures.

Unsafe automated operations.

Property destruction.

Civil liability may involve contractors, engineers, equipment manufacturers, software developers and operators, depending on the facts and applicable legal duties.

E. Digital commerce

AI agents increasingly negotiate contracts, process orders and execute automated transactions.

Disputes may arise where an AI agent:

Accepts an unauthorised transaction.

Misrepresents product information.

Executes an incorrect payment.

Violates contractual restrictions.

Produces misleading recommendations.

The legal status of the electronic transaction, authority of the person deploying the agent and allocation of contractual risk must be examined.

8. Evidentiary challenges in autonomous intelligence litigation

One of the most difficult issues is proving responsibility when the decision-making process is technically complex.

A. The black-box problem

Some AI systems cannot readily explain how they reached a particular conclusion.

This creates difficulties in identifying:

The original programming error.

The relevant training-data defect.

The model's internal decision process.

The moment a malfunction occurred.

The individual or organisation responsible.

B. Expert evidence

Courts may require specialist evidence concerning:

Algorithmic architecture.

Software defects.

Data integrity.

System performance.

Cybersecurity.

Technical causation.

Foreseeability of harm.

C. Digital audit trails

AI providers and operators should preserve:

System logs.

Decision histories.

Software versions.

Training and validation records.

Human intervention records.

Security incident reports.

Such records can assist in establishing the relationship between system behaviour and actual damage.

D. Burden of proof

The claimant generally needs to establish the applicable elements of liability and the connection between the harmful conduct and the loss.

Where information is exclusively controlled by an AI provider, access to technical evidence and appropriate expert examination may become particularly important.

9. Machine moral accountability versus human legal accountability

The distinction can be illustrated through the following examples.

SituationEthical issuePotentially responsible party
AI discriminates against a job applicantEquality and fairnessEmployer and potentially developer
Medical AI recommends dangerous treatmentPatient safetyHospital, professional and potentially provider
Autonomous vehicle causes an accidentPublic safetyOperator under applicable law
AI agent executes an unauthorised paymentFinancial integrityRelevant contracting parties and service providers
Industrial robot injures a workerHuman dignity and workplace safetyEmployer, operator or manufacturer
AI generates misleading financial adviceTransparency and honestyRelevant service provider or financial institution

The key principle is that an AI system's ability to make decisions does not necessarily give it the capacity to bear the legal consequences of those decisions.

10. Major challenges in the UAE legal framework

The development of autonomous intelligence raises several unresolved questions.

1. Absence of independent AI legal personality

There is no general UAE legal recognition of autonomous AI systems as independent civil legal persons capable of bearing their own civil liabilities.

2. Attribution of responsibility

Where developers, manufacturers and operators contribute to a single harmful outcome, determining the appropriate allocation of responsibility can be difficult.

3. Algorithmic opacity

Proprietary systems may prevent injured persons from obtaining sufficient information to establish causation.

4. Machine learning and unpredictable behaviour

An AI system may change its operational behaviour after deployment, making conventional fault analysis more complicated.

5. Cross-border liability

An AI platform may be developed in one country, hosted in another and operated by a UAE company.

This creates questions concerning jurisdiction, applicable law and enforcement.

6. Ethical standards versus enforceable obligations

Ethical AI guidelines provide important standards, but courts must distinguish between policy principles and legally enforceable duties.

7. Insurance and compensation

Traditional insurance policies may not adequately address losses resulting from autonomous decision-making, cyber incidents or interconnected machine failures.

11. Recommendations for strengthening UAE civil law

The following reforms could improve the legal treatment of autonomous intelligence.

Proposed reformLegal objective
AI civil liability legislationEstablish clear liability rules
Mandatory AI risk assessmentsPrevent foreseeable harm
Human oversight requirementsPreserve meaningful intervention
Algorithmic audit obligationsImprove evidentiary transparency
AI liability insuranceEnsure compensation
Developer accountability standardsEncourage safe design
AI incident reportingImprove public safety
Explainability requirementsFacilitate judicial review
Clear contractual allocationReduce disputes among stakeholders
Cross-border cooperationAddress international AI disputes

A particularly useful reform would be to introduce a risk-based liability framework distinguishing between low-risk AI applications and high-risk autonomous systems capable of causing serious physical, financial or personal harm.

12. Conclusion

Civil Law and Ethics of Autonomous Intelligence Systems and Machine Moral Accountability in the UAE represents an important development in modern civil jurisprudence.

The UAE already has several relevant legal foundations, including the Civil Transactions Law, Dubai's autonomous vehicle legislation, data protection legislation and national AI ethical governance principles.

The most important distinction is between machine autonomy and legal accountability.

An AI system may independently analyse information, generate decisions and perform complex operations. However, that technological independence does not automatically establish independent legal or moral responsibility.

The legal responsibility must be assessed through the conduct and obligations of the humans and organisations involved in developing, manufacturing, deploying, supervising and operating the system.

Central legal principle

Autonomous intelligence may exercise operational independence, but operational independence must not become a mechanism for escaping civil responsibility.

The future direction of UAE civil law should therefore combine technological innovation with human dignity, transparency, safety, fair treatment and effective compensation for persons harmed by autonomous systems.

13. Important legal provisions and authorities for further study

Legal sourceRelevance
Federal Law No. 5 of 1985, Civil Transactions LawTraditional civil liability framework
Federal Decree-Law No. 25 of 2025New Civil Transactions Law
Article 245 of the 2025 LawScope of harmful-act liability
Article 246 of the 2025 LawCompensation for harmful acts
Dubai Law No. 9 of 2023Autonomous vehicle regulation
Article 14 of Dubai Law No. 9 of 2023Operator's civil liability
Administrative Resolution No. 939 of 2025Autonomous vehicle implementing requirements
Federal Decree-Law No. 45 of 2021Personal Data Protection
UAE AI Ethics Principles and Guidelines (2022)Ethical AI governance
UAE Charter for the Development and Use of AIResponsible AI development and deployment

Research caution: The eight judicial authorities discussed above should not be described as eight UAE judgments establishing AI liability. They are a combination of DIFC judgments, a foreign judgment applying UAE law and comparative foreign authorities. The precise application of the 2025 Civil Transactions Law must also be checked against its commencement and transitional provisions when analysing a particular dispute.

 

 

LawGratis provides pro bono legal aid and resources to empower communities. Join us in making justice reachable for everyone! Follow our community on Instagram, YouTube, Facebook, and LinkedIn.

LEAVE A COMMENT