Civil Law and UAE Legal Personality of AI Agents — Debate

1. Introduction

The question whether an AI agent should have legal personality under UAE civil law is an emerging issue at the intersection of civil liability, contract law, agency, technology law and corporate law.

An AI agent is more than a conventional software programme when it can:

independently interpret instructions;

make decisions;

select tools;

communicate with third parties;

enter or initiate transactions;

control digital assets;

generate contractual communications;

perform tasks without continuous human intervention.

The central legal question is:

Should an autonomous AI agent itself become a legal person capable of holding rights, assuming obligations, owning assets, entering contracts and being sued?

Under the present UAE legal framework, the answer is no. UAE law recognises automated electronic transactions, including contracts formed without direct human intervention, but this does not give the automated system separate legal personality. Current UAE legal analysis treats AI systems as technological instruments whose legal consequences generally attach to the developer, owner, deployer, operator or other legally responsible person. (Global Practice Guides)

This creates an important distinction:

Autonomous action ≠ legal personality.

2. Meaning of Legal Personality

Legal personality means the law recognises an entity as capable of being the subject of legal rights and obligations.

A legal person may ordinarily be capable of:

owning property;

entering contracts;

acquiring rights;

assuming obligations;

bringing legal proceedings;

being sued;

being represented;

incurring liabilities;

possessing an independent legal estate.

Natural persons possess legal personality by operation of law.

Companies may acquire juristic/legal personality through incorporation or another legal mechanism.

The important question is whether an AI agent can occupy the third category.

3. AI Agent vs Legal Person

An AI agent should not be confused with:

a natural person;

a company;

a foundation;

a government entity;

an incorporated organisation.

An AI agent is presently a technological system.

For example:

Company A owns an AI purchasing agent.
The AI agent negotiates with Supplier B.
The agent autonomously accepts an offer for AED 500,000.

The legal question is not simply whether the AI made the acceptance.

The questions are:

Who owns the AI?

Who authorised its operation?

Who programmed or deployed it?

Who bears the financial consequences?

Was the agent acting within authority?

Was the third party entitled to rely on its conduct?

Who can be sued if something goes wrong?

Under current UAE principles, the answer generally leads back to a natural or legal person, rather than the AI system itself. (Global Practice Guides)

4. UAE Law Already Recognises Automated Contracting

This is one of the strongest arguments against confusing AI autonomy with legal personality.

The UAE's electronic-transactions legislation expressly recognises contracts formed through automated electronic systems.

Federal Decree-Law concerning Electronic Transactions and Trust Services provides that a contract may be made through automated electronic systems programmed for that purpose, and that the contract remains valid even without direct intervention by a natural person. It also addresses contracts between an automated electronic system belonging to a person and another person. (LittDB)

Therefore:

The UAE legal system can recognise the legal effect of autonomous electronic action without treating the machine as a legal person.

This is a crucial point in the debate.

5. The Core Legal Debate

There are two principal positions.

Position A — AI should NOT have legal personality

This is presently closer to the UAE legal position.

Arguments include:

AI lacks human legal capacity;

AI does not have an independent estate;

AI cannot meaningfully bear financial responsibility;

AI cannot be imprisoned or otherwise subjected to ordinary human sanctions;

liability can generally be traced to human or corporate actors;

legal personality could create accountability gaps;

victims might find it harder to obtain compensation.

A recent UAE-focused academic analysis concludes that giving autonomous AI systems legal personality is not presently a satisfactory solution because AI lacks features such as intention, legal capacity and independent assets necessary for effective responsibility. (mjsl.usim.edu.my)

6. Position B — AI Could Eventually Receive Limited Legal Personality

The opposing argument is based on technological autonomy.

An advanced AI agent may:

learn independently;

select among alternatives;

negotiate;

contract;

manage assets;

communicate;

execute transactions;

interact with other autonomous systems.

Supporters of limited personality could argue that law has previously created legal personality for entities that are not human beings.

A company is the obvious example.

Therefore, the argument is:

If law can create a legal person out of an organisational structure, why could it not create a limited legal person around a sufficiently autonomous AI system?

However, this would require legislation rather than simply judicial recognition.

7. The Company Analogy

Corporate personality provides the strongest conceptual analogy.

A UAE company can have:

separate property;

separate liabilities;

contractual capacity;

litigation capacity;

representatives;

independent existence from shareholders.

A DIFC judgment recently reiterated the UAE corporate principle that a company has legal personality separate from its owners and that a subsidiary has independent legal personality and financial liabilities. Normand v Nathaniel [2024] DIFC SCT 125 is useful in this respect. (DIFC Courts)

But the analogy has a fundamental limitation.

A company is a legal person because law deliberately creates it as one.

An AI system does not automatically become a legal person merely because it is technologically sophisticated.

Thus:

Corporate personality is legally constructed; AI autonomy is technologically constructed.

8. Case Law 1 — Normand v Nathaniel [2024] DIFC SCT 125

Principle

The DIFC Small Claims Tribunal discussed the separate legal personality of companies and subsidiaries.

The judgment recognised that a company has an identity separate from its owners and can possess its own rights and liabilities. (DIFC Courts)

Relevance to AI

This provides a useful test for AI personality.

If AI were ever granted legal personality, legislation would need to establish:

the AI's separate estate;

its legal identity;

its ability to contract;

its liability;

representation;

ownership of assets;

mechanisms for dissolution or termination.

Debate

Pro-AI personality:
The corporate model demonstrates that legal personality need not be limited to human beings.

Against AI personality:
The company receives personality through legislation and possesses an identifiable legal estate. Ordinary AI systems presently lack those characteristics.

9. Case Law 2 — Currency Matters Middle East v Michael Page International Ltd [2018] DIFC CFI 039

This case concerned apparent authority.

The DIFC Court considered whether an individual who lacked actual authority could nevertheless bind a company because the company's conduct created the appearance of authority.

The Court explained that apparent authority may arise where the principal's conduct causes a third party reasonably to believe that an agent has authority. (DIFC Courts)

Relevance to AI

This doctrine provides a potentially powerful framework for autonomous AI agents.

Suppose:

Company A deploys an AI purchasing agent on its website.

A supplier reasonably believes that the AI has authority to purchase goods.

The AI concludes a contract.

The question becomes:

Should Company A be bound because it created the appearance that the AI was authorised to contract?

The agency model provides a solution without creating AI personality.

10. Case Law 3 — International Electro-Mechanical Services Co. LLC v Emirates Speciality Hospital FZ-LLC [2020] DIFC CFI 114

The DIFC Court considered actual, implied and apparent authority.

The judgment referred to UAE jurisprudence recognising that agency may be implied from circumstances and discussed Dubai Court of Cassation authorities concerning apparent authority. (DIFC Courts)

Relevance to AI agents

The case supports an important alternative to AI legal personality:

Treat the AI as an instrument through which an existing legal person exercises authority.

For example:

Principal: Company
↓
AI Agent: Automated intermediary
↓
Third Party: Customer

The AI does not become the principal.

Instead, the company remains legally responsible for acts performed within the AI's authorised functions.

11. Case Law 4 — Khaled Salem Musabeh Humad Al Mheiri v John Cameron [2025] DIFC CA 008

This recent DIFC Court of Appeal decision examined ostensible authority under UAE law.

The Court discussed circumstances in which a principal may become responsible where:

the agent acts in the principal's name;

the third party acts in good faith;

the principal's conduct creates a reasonable impression that authority exists. (DIFC Courts)

Relevance to AI

This is highly relevant to autonomous AI.

Imagine that an AI agent:

uses the company's identity;

communicates through its systems;

negotiates with customers;

signs electronically;

performs transactions.

A third party may reasonably believe that the AI is authorised.

The legal problem can therefore be addressed through agency and attribution, rather than personality.

12. Case Law 5 — Alarabi Investments Ltd v Cron AI Ltd [2026] DIFC CFI 030/2025

This is especially relevant because the defendant itself was an AI-related company.

The DIFC Court dealt with proceedings involving Cron AI Ltd, a corporate defendant. The case demonstrates an important distinction:

The legal person is the company developing or operating AI—not the AI technology itself.

The court record identifies Cron AI Ltd as the defendant in the proceedings. (DIFC Courts)

Significance

The case does not establish AI legal personality.

Instead, it illustrates the present legal architecture:

AI business/company → legal personality

rather than:

AI system → legal personality

This distinction is fundamental to the current UAE debate.

13. Case Law 6 — Klesta Eshja & Hair Creators Salon LLC v Salah Masri & Others [2025] DIFC CFI 066/2024

This is one of the most significant recent UAE cases concerning AI's role in litigation.

The defendants' pleadings had allegedly been prepared substantially with AI assistance and contained false case references and misleading material.

The Court ordered the relevant defences to be struck out and imposed costs consequences. The claimants later submitted a substantial statement of costs. (DIFC Courts)

Relevance to AI personality

The case demonstrates that the judicial system treats AI as a tool used by litigants, rather than as an independent legal actor.

The AI did not:

become a party;

possess standing;

appear as defendant;

assume costs liability.

The human/legal persons using the technology remained responsible.

14. Case Law 7 — VTB Bank PJSC v Kuanyshev & Others [2026] DIFC CFI 121/2025

This case is highly significant for AI governance.

The Court identified numerous apparently nonexistent or incorrect authorities in the defendants' submissions and expressly raised the possibility that the material had been generated by AI. The judgment referred to the DIFC's guidance concerning the use of large language models and generative AI in proceedings. (DIFC Courts)

Legal significance

The case reinforces a fundamental principle:

Using AI does not transfer legal responsibility from the litigant to the AI.

The human party remains responsible for:

accuracy;

verification;

pleadings;

citations;

representations to the court.

This is directly relevant to the debate over AI legal personality.

15. What These Seven Cases Demonstrate

16. The Attribution Model

The most practical current UAE model is:

AI Agent → acts → legal person → legal consequences

For example:

Company

↓ owns/deploys

AI agent

↓ autonomously negotiates

Customer

↓ enters transaction

Contract

The contract can be valid without the AI becoming a legal person because UAE electronic-transactions law already recognises automated contracting. (LittDB)

17. The Personality Model

The alternative would be:

Company

↓ creates

AI legal person

↓ owns assets

AI agent

↓ contracts

Third party

In this model, the AI would potentially possess:

separate assets;

separate liabilities;

legal standing;

contractual capacity;

litigation capacity.

But the UAE would have to legislate a new category of legal person.

18. Why Legal Personality Might Be Attractive

There are several arguments in favour.

A. Autonomous decision-making

Highly autonomous systems can make decisions without immediate human intervention.

B. Clear allocation of liability

A legally recognised AI entity could potentially provide a single liability point.

C. Technological neutrality

Legal personality could theoretically adapt the law to increasingly autonomous technologies.

D. Separate asset pool

An AI entity could potentially maintain an asset or insurance fund for liabilities.

E. Commercial efficiency

AI agents could theoretically operate in markets without requiring constant human authorisation.

19. Arguments Against AI Legal Personality

A. Lack of independent assets

If the AI owns no property, legal personality may provide little practical compensation.

B. Lack of meaningful intention

An AI can generate decisions but this does not necessarily equate to legally recognised intention.

C. Accountability problem

Giving AI personality could make developers and deployers attempt to shift responsibility to the machine.

D. Victim protection

A victim may obtain a judgment against an AI but still have no meaningful source of compensation.

E. Artificial personality may be unnecessary

Agency and attribution doctrines can already allocate responsibility.

F. Regulatory complexity

A new category would require rules concerning:

incorporation;

registration;

taxation;

insurance;

assets;

insolvency;

dissolution;

representation;

liability.

20. The “Electronic Agent” Problem

The expression “electronic agent” can be misleading.

An electronic system may function as an agent in the technological sense without becoming an agent in the legal sense.

UAE electronic-transactions law allows automated systems to form contracts. (LittDB)

But this does not mean:

Automated system = legal person.

The better interpretation is:

Automated system = mechanism through which a legal person performs an electronically recognised transaction.

21. AI and Civil Liability

Suppose an autonomous AI agent causes AED 10 million in damage.

Possible responsible parties include:

Developer

If defective design caused the harm.

Owner

If the system was inadequately maintained.

Deployer

If the system was used negligently.

Operator

If the operator improperly configured the system.

User

If the user intentionally or negligently instructed the system to perform harmful activity.

AI itself

Under current UAE law:

No separate legal personality.

Current UAE AI analysis similarly identifies the developer, deployer or user as the persons to whom liability must generally be attributed. (Global Practice Guides)

22. AI and Contractual Liability

Suppose an AI agent orders 1,000 products without human confirmation.

The seller argues:

“The AI made the contract.”

The buyer argues:

“No human authorised it.”

UAE electronic-transactions law already recognises contracts formed by automated electronic systems in appropriate circumstances. (LittDB)

The legal questions therefore become:

Did the system belong to or operate for the relevant person?

Was the system programmed for that transaction?

Was the counterparty aware, or should it have been aware, of automated operation?

Was the AI operating within the relevant authority?

Was there fraud or misuse?

What does the underlying contract say?

These questions can be answered without granting personality to the AI.

23. AI and Agency Law

Agency may become one of the most important legal mechanisms for autonomous AI.

Traditional agency

Principal → Human agent → Third party

AI agency

Principal → AI system → Third party

The AI is technically performing the intermediary function.

But legally:

Principal → legal attribution → third party

This structure preserves human/corporate accountability.

24. AI and Corporate Personality

A company can therefore be viewed as the legal container for AI activity.

For example:

AI company

owns AI model;

employs developers;

holds bank accounts;

owns intellectual property;

enters contracts;

maintains insurance;

pays liabilities.

AI agent

executes transactions;

communicates;

analyses information;

interacts with APIs;

makes operational decisions.

The company remains the legal person.

This is consistent with the separate-personality principle discussed in Normand v Nathaniel. (DIFC Courts)

25. AI Legal Personality and the New Civil Transactions Law

The new UAE Civil Transactions Law, effective from 1 June 2026, modernises the general civil-law framework and emphasises contemporary transactional realities, legal capacity and freedom of contract. (UAE Legislation)

However, modernisation of civil transactions should not be interpreted as automatically creating AI legal personality.

A distinction must be maintained between:

Legal modernisation

and

Creation of a new legal person.

The latter requires a clear legal basis.

26. Could UAE Law Create “Limited AI Personality”?

A possible future model could provide AI with limited legal personality.

For example:

AI Entity A

Could:

own a designated liability fund;

enter specified contracts;

have mandatory insurance;

have a registered human/legal-person custodian;

operate within predefined limits;

be subject to regulatory supervision.

But it might not:

possess unrestricted property rights;

exercise political rights;

make personal-status decisions;

escape human responsibility;

operate without a responsible custodian.

This would be closer to special-purpose juristic personality than human-like personality.

27. The “Electronic Person” Model

Another theoretical possibility is a new statutory category:

Electronic Legal Person

Such a system might require:

Government registration.

Identifiable owner/custodian.

Mandatory insurance.

Minimum capital or liability fund.

Digital identity.

Audit logs.

Human oversight.

Cybersecurity standards.

Defined scope of legal capacity.

Rules for termination and insolvency.

Without these safeguards, personality could create more uncertainty than it solves.

28. AI Legal Personality and Tort Law

The tort problem is particularly difficult.

Traditional civil liability normally asks:

Who committed the wrongful act?

With AI:

The AI performed the immediate operation.

But the legal system can instead ask:

Who created, deployed, controlled or failed to supervise the system?

This allows liability to be connected to human or corporate responsibility.

The UAE-focused literature on autonomous AI similarly questions whether traditional fault and causation principles can deal adequately with autonomous systems, while concluding that simply granting AI legal personality does not necessarily solve the compensation problem. (mjsl.usim.edu.my)

29. AI and Causation

Autonomous AI creates complex causation questions.

Example:

Developer → Model → AI agent → API → autonomous decision → financial loss.

There may be several potential causes.

A court could need to determine:

Was the software defective?

Was deployment negligent?

Was the user negligent?

Was the third party contributor to the harm?

Was the result foreseeable?

Did an independent intervening event occur?

Legal personality of AI does not automatically solve these causation problems.

30. AI and Evidence

AI agents create extensive evidence:

logs;

prompts;

outputs;

API records;

timestamps;

system instructions;

transaction histories;

audit trails.

The UAE Evidence Law already provides a framework for electronic evidence.

Therefore, the modern legal problem is increasingly:

How do we prove what an AI agent actually did?

rather than:

Does the AI have a personality?

31. AI and Digital Identity

If AI agents become commercially autonomous, legal systems may need reliable identification.

An AI agent could potentially possess:

digital credentials;

cryptographic keys;

registered identifiers;

API credentials;

transaction permissions.

But possession of a digital identity does not itself create legal personality.

This is analogous to:

Having a bank account or electronic signature ≠ becoming a legal person.

32. AI and Smart Contracts

AI agents may interact with smart contracts.

For example:

AI agent identifies market condition → executes blockchain transaction → smart contract transfers digital asset.

Three separate legal questions arise:

Is the AI authorised?

Is the underlying transaction legally valid?

Who bears liability if the AI makes a mistake?

Again, AI personality is not necessarily required.

33. AI and Consumer Protection

Imagine an AI shopping agent purchases a defective product.

The consumer should not lose legal protection simply because:

“The purchase was made by an AI.”

A consumer-protection system should identify:

the consumer;

supplier;

platform;

AI provider;

deployer;

contractual relationship.

This strongly supports an attribution model rather than unrestricted AI personality.

34. AI and Data Protection

An AI agent may process:

names;

financial information;

location data;

behavioural information;

identification information.

The relevant obligations should remain attached to the legally responsible entity.

Giving the AI personality should not allow an organisation to argue:

“The AI processed the data, therefore we are not responsible.”

That would undermine data-protection principles.

35. AI and Intellectual Property

Another difficult question is ownership of AI-generated material.

Current UAE legal analysis does not treat the AI itself as a legal author possessing personality merely because it generated content. The relevant legal framework instead focuses on human/legal persons and applicable copyright principles. (Global Practice Guides)

Thus:

AI generation does not automatically imply AI ownership.

36. AI and Court Proceedings

The recent DIFC cases demonstrate an important principle.

Klesta Eshja

AI-assisted pleadings containing false material resulted in procedural and cost consequences for the parties. (DIFC Courts)

VTB Bank

The Court considered the possibility that inaccurate authorities had been generated by AI and referred to guidance governing generative-AI use in proceedings. (DIFC Courts)

These cases reinforce:

Human users remain accountable for AI-assisted legal activity.

37. Arguments for a UAE “AI Custodian” Model

Instead of legal personality, UAE law could develop a custodian model.

Structure

AI system

↓ controlled by

AI custodian

↓ responsible for

compliance;

insurance;

cybersecurity;

supervision;

records;

compensation.

This would preserve the advantages of autonomy while maintaining identifiable human/legal responsibility.

Academic work specifically examining autonomous AI under UAE law has proposed a custodian-oriented framework as an alternative to granting AI independent legal personality. (MDPI)

38. Comparative Debate

39. Six Central Legal Questions for UAE Courts

If an AI-agent dispute reaches a UAE court, the court may need to determine:

Question 1

Who owns or controls the AI?

Question 2

Who authorised the AI's operation?

Question 3

Was the AI acting within its assigned authority?

Question 4

Could the third party reasonably rely upon its apparent authority?

Question 5

Who caused or contributed to the damage?

Question 6

What contractual, tortious, electronic-transaction or regulatory rule applies?

These questions can be answered without recognising the AI as a separate legal person.

40. Major Legal Policy Debate

The debate can therefore be summarised in two competing propositions.

Proposition 1 — Personality is unnecessary

AI should remain a sophisticated instrument. Existing principles of contract, agency, tort, electronic transactions, corporate liability and evidence can allocate responsibility.

Proposition 2 — Personality may eventually become necessary

As AI agents become increasingly autonomous and economically independent, traditional attribution may become inadequate, requiring a new form of limited legal personality.

At present, the first model is much closer to UAE law.

41. Recommended UAE Legal Framework for Autonomous AI

A practical future framework could combine:

1. Mandatory AI registration

For high-risk autonomous commercial agents.

2. Human/legal-person custodian

Every high-risk AI agent should have an identifiable responsible person or company.

3. Mandatory insurance

Insurance could provide compensation for AI-related civil damage.

4. Auditability

AI agents should maintain transaction logs.

5. Digital identity

Every autonomous commercial agent could possess a traceable identifier.

6. Authority limits

The AI's transaction limits should be technically enforceable.

7. Human override

A responsible person should be able to suspend the system.

8. Liability rules

Law should allocate responsibility among developer, owner, deployer and user.

9. Consumer protection

AI autonomy should not reduce consumer rights.

10. Court supervision

Courts should retain authority to examine AI-generated evidence and conduct.

42. Examination Answer Structure

For an examination question:

“Discuss whether AI agents should possess legal personality under UAE civil law.”

Use this structure:

Introduction

Define AI agents and legal personality.

Existing UAE position

Explain that AI currently has no independent legal personality.

Automated transactions

Discuss UAE recognition of automated electronic contracts.

Agency

Explain actual, implied and apparent authority.

Corporate personality

Compare AI with companies.

Civil liability

Discuss developer, owner, deployer and operator liability.

Evidence

Discuss electronic evidence and AI-generated records.

Case law

Use at least six authorities.

Arguments for personality

Discuss autonomy, accountability and commercial efficiency.

Arguments against personality

Discuss assets, capacity, responsibility and victim protection.

Future reform

Discuss an AI-custodian or limited-personality model.

Conclusion

Explain that present UAE law favours attribution to human/legal persons, while future legislation may develop a specialised framework for highly autonomous AI.

43. Quick Revision Chart

44. Conclusion

The debate over legal personality of AI agents in UAE civil law should begin with an important distinction:

The UAE already recognises legal consequences of automated electronic action, but it does not thereby recognise the AI system itself as a legal person. (LittDB)

The existing legal architecture can generally operate through:

Principal / owner / developer / deployer
↓
AI agent
↓
Automated action
↓
Legal consequences attributed to the responsible person or entity

The UAE's modern civil-law framework, electronic-transactions legislation and developing digital-justice environment make this attribution model increasingly important. The recent DIFC cases involving Cron AI, AI-assisted pleadings and potentially AI-generated false authorities demonstrate that UAE courts are already confronting practical AI questions, but these cases do not recognise AI itself as a legal person. (DIFC Courts)

Key Cases for Revision

Normand v Nathaniel [2024] DIFC SCT 125 — separate corporate legal personality.

Currency Matters Middle East v Michael Page International Ltd [2018] DIFC CFI 039 — apparent authority.

International Electro-Mechanical Services Co. LLC v Emirates Speciality Hospital FZ-LLC [2020] DIFC CFI 114 — actual, implied and apparent authority.

Khaled Salem Musabeh Humad Al Mheiri v John Cameron [2025] DIFC CA 008 — ostensible authority under UAE law.

Alarabi Investments Ltd v Cron AI Ltd [2026] DIFC CFI 030/2025 — AI-related company as legal defendant, illustrating the distinction between AI technology and the legal entity operating it.

Klesta Eshja & Hair Creators Salon LLC v Salah Masri & Others [2025] DIFC CFI 066/2024 — consequences of AI-assisted litigation material.

VTB Bank PJSC v Kuanyshev & Others [2026] DIFC CFI 121/2025 — AI-generated legal material and continuing responsibility of litigants.

Core principle:

Under present UAE civil law, an AI agent may act autonomously and may generate legally effective transactions, but autonomy does not itself create legal personality. The stronger current model is to attribute the AI's conduct to the legally responsible human or corporate actor, while future legislation could potentially introduce a limited personality or AI-custodian regime for highly autonomous systems.