Civil Law and UAE: Machine Agency and Legal Subjectivity Debates

1. Introduction

Machine agency concerns whether an automated system, AI system, robot, algorithm, smart contract, or other computational system can be treated as an agent capable of producing legally relevant acts.

Legal subjectivity is a different and much broader question:

Can the machine itself be recognised as a legal subject possessing rights, duties, liabilities, property, legal capacity, or the ability to sue and be sued?

Under current UAE law, these questions should be kept separate.

The UAE has expressly recognised automated electronic transactions, but that does not amount to recognition of AI or machines as independent legal persons. Federal Decree-Law No. 46 of 2021 provides that contracts may be formed between automated electronic mediums programmed for that purpose and that such contracts can be valid, enforceable and legally effective.

At the same time, the current Civil Transactions Law, Federal Decree by Law No. 25 of 2025, effective from 1 June 2026, defines the categories of legal persons. Article 92 includes the State, public authorities, companies, associations, foundations, waqfs and other groups of persons or property to which the law grants legal personality. Article 93 provides for the rights of a legal person, while Article 94 provides for its independent patrimony, legal capacity, right to litigate and domicile.

Machines and AI systems are not included as a separate category of legal person.

Therefore, the present UAE approach can be understood as:

Machine may perform legally relevant functions without itself becoming a legal person.

2. What Is Machine Agency?

Traditional agency involves three participants:

Principal → Agent → Third Party

For example:

Company A authorises Employee B to conclude a contract with Company C.

The employee acts as an agent for the company.

Machine agency creates a different structure:

Human/Company → Automated System → Third Party

For example:

A company programs an automated trading system to buy securities when the price reaches a specified level.

The system may independently:

• identify the condition;
• make the calculation;
• communicate acceptance;
• place an order;
• execute payment;
• generate records.

The difficult legal question is:

Who legally acted when the machine generated the transaction?

3. Machine Agency Is Not Machine Personhood

This is the most important distinction.

Machine agency

The machine is an instrument through which a human or legal entity acts.

Machine personhood

The machine itself becomes a legal subject.

These are not equivalent.

For example:

A company programs an AI purchasing system to order raw materials automatically.

If the system places an order, the legal analysis may attribute the transaction to the company under the applicable electronic-transactions framework.

It does not follow that:

“The AI bought the materials.”

in the sense that the AI has become an independent legal person.

4. UAE Statutory Foundation: Automated Electronic Transactions

Federal Decree-Law No. 46 of 2021 is particularly important.

Article 10

Electronic offer and acceptance can form a contract, and a contract does not lose validity or enforceability merely because it exists electronically.

Article 11

The legislation specifically recognises contracts formed between automated electronic mediums comprising electronic information systems programmed for that purpose. Such contracts may be valid, enforceable and legally effective.

This is a significant legislative choice.

The law effectively says:

A human being does not have to manually press every button for a legally valid electronic transaction to occur.

But the statute does not say:

The automated system itself becomes a legal person.

5. Why Article 11 Matters

Article 11 solves one problem while leaving another unresolved.

Problem solved

Can an automated system participate in contract formation?

Yes.

Problem remaining

Does the automated system itself possess:

• legal personality?
• independent assets?
• legal capacity?
• liability?
• standing to sue?
• independent obligations?

The legislation does not establish machines as a separate category of legal person.

This distinction is fundamental to the machine-agency debate.

6. Current Civil Transactions Law and Legal Personality

The new UAE Civil Transactions Law is particularly relevant.

Article 92 identifies legal persons, including:

1. the State and Emirates;
2. public authorities and institutions;
3. recognised religious bodies;
4. waqfs;
5. civil and commercial companies;
6. associations and foundations;
7. other groups of persons or property to which legal personality is granted by law. 

Article 94 further provides that a legal person has:

• independent financial patrimony;
• legal capacity within legally prescribed limits;
• right to litigate;
• independent domicile. 

This gives us a useful test.

A machine does not become a legal person merely because it:

• makes decisions;
• communicates;
• learns;
• predicts;
• acts autonomously;
• interacts with humans;
• controls digital assets.

Legal personality requires a legal foundation.

7. Four Possible Models of Machine Legal Status

The debate can be divided into four models.

Model 1 — Machine as Tool

The machine is merely an instrument.

Human/company → machine → transaction

Liability normally remains with the relevant human or legal entity.

This is closest to the present UAE statutory structure.

Model 2 — Machine as Automated Agent

The machine performs agency-like functions.

It may:

• negotiate;
• communicate;
• accept;
• reject;
• execute;
• purchase;
• sell.

But legal consequences are attributed to the human or legal person responsible for its deployment.

Model 3 — Limited Electronic Personality

Under a future legislative model, a machine could theoretically receive limited legal capacity for specific purposes.

For example:

• holding limited assets;
• paying specified obligations;
• maintaining an automated contractual account.

This would require legislation defining:

• capital;
• liability;
• representation;
• registration;
• taxation;
• insolvency;
• insurance.

The present UAE Civil Transactions Law does not establish such a general category for machines.

Model 4 — Full Machine Legal Personhood

The strongest theory would treat an advanced autonomous AI as an independent legal subject.

It could theoretically:

• own property;
• contract;
• sue;
• be sued;
• incur debts;
• commit civil wrongs;
• possess a separate patrimony.

This would represent a major change in private-law theory.

There is presently no general UAE civil-law rule granting AI systems this status.

8. Case Law 1 — ICICI Bank Ltd v Bavaguthu Raghuram Shetty [2022] DIFC CFI 034

This large financial dispute illustrates an important principle for machine agency: complex electronic and financial transactions remain disputes between legally recognised parties.

The proceedings involved extensive financial documentation, procedural orders and judicial examination rather than treating technological or automated records as independent legal actors.

Relevance

A machine can generate:

• transaction records;
• payment instructions;
• financial communications;
• automated calculations.

But the legal dispute remains centred on the rights and obligations of the recognised parties.

Principle

Technological execution does not automatically create a new legal subject.

9. Case Law 2 — Graciela Limited v Giacobbe [2014] DIFC CFI 027

This is one of the most useful technology-related authorities.

The defendant was found responsible for deliberately interfering with the claimant's IT system. The court examined technical and circumstantial evidence concerning the attack and awarded USD 690,533 in compensatory damages.

The damages included:

• system restoration;
• investigation;
• emergency servers;
• employee time;
• disruption-related consequences.

Relevance to machine agency

The IT system was the object of interference, but the legal responsibility was attributed to the human defendant.

The computer system itself was not treated as:

perpetrator + legal person + independent defendant.

Principle

An autonomous or technically complex system does not automatically become the bearer of legal responsibility.

10. Case Law 3 — Linux v Lizeth [2022] DIFC SCT 237

This dispute concerned a Software Development Agreement for development of an e-commerce and restaurant-management platform.

The claimant alleged that the defendant had delivered a third-party platform rather than the originally promised software.

The case demonstrates an important legal distinction:

The software may be the subject matter of a contract, but the contracting parties remain legally responsible.

The platform was not treated as a contracting party.

Principle

Software functionality does not itself establish legal personality.

11. Case Law 4 — Alucor Ltd v Rohr Rein Chemie Middle East LLC [2021] DIFC TCD 001

This Technology and Construction Division case concerned a subcontract for work on the Al Taweelah Alumina Refinery Project in Abu Dhabi.

The dispute involved contractual obligations, jurisdiction and proceedings in the Dubai courts and DIFC Courts.

Relevance

Large technological and engineering projects involve:

• automated systems;
• engineering software;
• digital records;
• electronic communications;
• technical calculations.

Yet legal responsibility continues to attach to the contractual parties.

Principle

Technical autonomy within a project does not automatically create independent legal subjectivity.

12. Case Law 5 — Techteryx Ltd v Aria Commodities DMCC & Others [2025] DIFC DEC 001

This is particularly important for modern digital assets.

The DIFC Digital Economy Court dealt with a dispute concerning approximately USD 456 million in reserves associated with TrueUSD. The court granted proprietary and worldwide freezing relief and required extensive disclosure concerning the funds and traceable proceeds.

The Digital Economy Court continues to issue orders in the matter in 2026.

Relevance

The dispute demonstrates that:

• digital assets can be treated as legally significant property;
• digital systems can be the subject of judicial orders;
• courts can control digital assets through human/legal intermediaries.

But the digital system or token infrastructure was not thereby transformed into an independent legal person.

Principle

Digital autonomy and digital property are not synonymous with legal personality.

13. Case Law 6 — Panther Real Estate Development LLC v Modern Executive Systems Contracting LLC [2022] DIFC CA 016

This DIFC Court of Appeal case concerned a technically complex construction dispute. The Technology and Construction Division judgment was appealed to the Court of Appeal.

The case demonstrates how highly technical contractual environments remain governed by:

• contractual obligations;
• legal interpretation;
• evidence;
• judicial determination.

Relevance

An engineering system may independently:

• calculate quantities;
• generate measurements;
• identify delays;
• produce schedules.

But the system's output does not itself determine the legal rights of the contracting parties.

Principle

Technical decision-making is not necessarily legal decision-making.

14. Case Law 7 — Ondina v Olin [2025] DIFC CFI 046

This case arose from an SCT proceeding and subsequent appeal process concerning a contractual dispute. The DIFC Court of First Instance dismissed the appeal and dealt with the procedural questions under the DIFC Rules.

Although this is not a machine-personhood judgment, it illustrates an important methodological point:

The legal system determines the status and consequences of transactions through recognised procedural and substantive rules rather than through the technology alone.

It is therefore useful as an adjacent authority when discussing the distinction between automated activity and legal adjudication.

15. Case Law 8 — Alarabi Investments Ltd v Cron AI Ltd [2026] DIFC CFI 030

This is a particularly relevant modern case because the defendant itself was an AI-related company.

The case involved a default judgment and a later application seeking to set it aside. The procedural history demonstrates that the legal defendant was Cron AI Ltd, a corporate entity, rather than an AI system being treated as an independent legal person.

Importance

This provides a useful real-world distinction:

AI company ≠ AI legal person.

A company developing or using AI can be a legal person under ordinary corporate law.

The AI technology itself is not thereby granted separate legal personality.

16. Case Law 9 — Klesta Eshja & Hair Creators Salon LLC v Salah Masri & Others [2025/2026] DIFC CFI 066/2024

This case involved the use of artificial intelligence in preparing court pleadings.

The defendants' amended defences were found to have been substantially prepared with AI assistance and contained false references and misleading material. The court struck out the amended defences and made costs orders arising from the improper use of AI.

Importance for machine agency

The court did not treat the AI system as:

• the lawyer;
• the litigant;
• the responsible legal representative.

Instead, responsibility remained with the human parties and their litigation conduct.

Principle

AI assistance does not transfer procedural responsibility from the human/legal actor to the machine.

17. Case Law 10 — Stelian Gheorghe v BSA Ahmad Bin Hezeem & Associates LLP [2025] DIFC CFI 045

The case involved allegations that claim documents and evidence may have been generated partly using AI. The court noted that errors of law have no place in witness evidence filed by lawyers and ultimately ordered a stay in favour of arbitration.

Importance

This illustrates another dimension of machine agency:

A lawyer cannot simply say:

“The AI produced it.”

and thereby eliminate professional or procedural responsibility.

The legal actor remains accountable for material filed with the court.

18. Why These Cases Matter

The cases collectively demonstrate an emerging principle:

19. The “Electronic Agent” Debate

There are three major theories.

Theory A — Instrument Theory

The machine is simply a tool.

Example:

A company programs software to purchase goods automatically.

The company remains the principal legal actor.

This fits comfortably with the UAE's statutory recognition of automated electronic transactions.

Theory B — Functional Agency Theory

The machine performs many functions traditionally associated with an agent.

It may:

• communicate;
• negotiate;
• select;
• contract;
• execute.

But it does not possess independent legal personality.

The law effectively attributes the machine's conduct to the relevant human/legal actor.

Theory C — Autonomous Legal Agent

Under this theory, highly autonomous AI could eventually be recognised as an independent legal agent.

It might have:

• independent decision-making;
• separate assets;
• limited liability;
• contractual capacity;
• procedural standing.

This would require significant legislative development.

20. Why Legal Subjectivity Is More Difficult Than Agency

Agency concerns:

Who is acting for whom?

Legal subjectivity asks:

Who is the holder of the right or obligation?

This is a much deeper question.

Suppose an autonomous AI enters a contract.

Three possibilities exist:

Option 1

The developer is responsible.

Option 2

The user/operator is responsible.

Option 3

The AI itself is responsible.

The third option requires an answer to:

Where does the AI's liability come from?

21. The Patrimony Problem

One of the strongest objections to machine personhood is the patrimony problem.

A legal person generally has an independent financial patrimony.

Under Article 94 of the current UAE Civil Transactions Law, a legal person has an independent financial patrimony and legal capacity.

Suppose an AI causes AED 5 million of damage.

If the AI is a legal person:

What assets satisfy the judgment?

Possible answers might be:

• AI-owned cryptocurrency;
• dedicated capital;
• insurance;
• developer-funded account.

But without a legally recognised patrimony, machine personality would have limited practical value.

22. The Liability Problem

Consider an autonomous vehicle.

The system makes a decision that causes an accident.

Potentially responsible parties include:

1. manufacturer;
2. software developer;
3. owner;
4. operator;
5. maintenance provider;
6. data provider;
7. insurer;
8. autonomous system.

If the machine itself becomes a legal person, another question arises:

Does machine personality merely create another defendant while leaving the human actors responsible?

If so, the economic value of machine personality becomes uncertain.

23. The Fault Problem

Traditional civil liability frequently examines:

• intention;
• negligence;
• knowledge;
• breach of duty;
• foreseeability;
• causation.

Can an AI:

• intend?
• know?
• be negligent?
• act in bad faith?

Philosophical theories may answer these questions differently.

But legal attribution is ultimately a matter of legislative and judicial rules, not merely technological capability.

24. The Responsibility-Gap Problem

Suppose:

Developer says: “The user controlled the AI.”

User says: “The developer designed the AI.”

Manufacturer says: “The software made the decision.”

AI says nothing because it has no recognised legal personality.

This creates a potential responsibility gap.

Civil law must therefore determine:

Which human or legal person bears responsibility for designing, deploying, controlling or benefiting from the system?

25. Machine Agency and Contract Formation

A particularly important UAE distinction is:

Traditional contract

Human A → Human B

Automated contract

Human/company → automated system → automated system → human/company

Federal Decree-Law No. 46 of 2021 expressly accommodates the second structure. Article 11 recognises contracts between automated electronic mediums.

Therefore, the law does not require:

human hand + human hand = valid contract

The parties can use automated systems.

But this is functional recognition, not legal personhood.

26. Smart Contracts

Smart contracts provide a powerful example.

A smart contract can automatically:

• receive payment;
• verify conditions;
• transfer tokens;
• release collateral;
• impose a programmed consequence.

The code may execute without human intervention.

But legal questions remain:

• Was there a valid underlying agreement?
• Who owned the asset?
• Who programmed the system?
• Who controlled the wallet?
• Was there fraud?
• Was there mistake?
• What happens if the code produces an unintended result?

The Techteryx proceedings demonstrate how the DIFC Digital Economy Court can apply conventional proprietary and injunctive remedies in a dispute involving digital assets.

27. Machine Agency and Evidence

A machine can generate:

• logs;
• timestamps;
• transaction records;
• blockchain entries;
• audit trails;
• AI outputs.

But evidence is not automatically equivalent to truth.

The court may still ask:

1. Is the record authentic?
2. Was it altered?
3. What system generated it?
4. Who controlled the system?
5. Was the system functioning correctly?
6. Is there corroborating evidence?

Graciela v Giacobbe demonstrates how technical evidence concerning an IT system can be examined to establish human responsibility.

28. Machine Agency and Judicial AI

The DIFC has gone further institutionally than simply recognising digital transactions.

Part 58 of the DIFC Rules establishes the Digital Economy Court and expressly includes disputes involving:

• artificial intelligence;
• digital assets;
• blockchain;
• fintech;
• complex databases;
• cloud data;
• e-commerce;
• digital payment platforms;
• automatic dispute-resolution processes. 

The rules also permit AI-driven smart forms and decision-tree software to assist with the conduct and disposal of certain digital-economy claims.

But this does not mean the AI system becomes the judge.

The court remains the legal institution exercising judicial authority.

29. Human-in-the-Loop Principle

A useful modern civil-law model is:

Machine performs → Human/legal person assumes responsibility → Court reviews legality.

This preserves:

• accountability;
• procedural fairness;
• legal personality;
• judicial supervision.

It avoids the proposition that:

“The machine decided, therefore nobody is legally responsible.”

30. AI-Assisted Legal Practice

Recent DIFC decisions show the practical importance of this issue.

In Klesta Eshja, AI-generated false authorities resulted in serious procedural consequences, including striking out of the amended defences and costs consequences.

The broader lesson is:

Delegating a task to AI does not necessarily delegate the legal responsibility associated with that task.

This is particularly important for:

• lawyers;
• accountants;
• financial institutions;
• compliance officers;
• corporate directors;
• expert witnesses.

31. Machine Agency and Corporate Law

A corporation is already a legal person.

Therefore, a useful structure is:

Company → AI system → transaction

The company can use the AI as an operational mechanism without granting the AI independent personality.

This is similar in conceptual terms to:

Company → employee → transaction

The difference is that an employee has independent human legal personality, whereas an automated system does not automatically have it.

32. Machine Agency and Attribution

Attribution can potentially depend upon:

• who programmed the machine;
• who authorised its deployment;
• who controlled it;
• who benefited from its operation;
• what contractual arrangements existed;
• whether the system acted within programmed parameters;
• whether human intervention occurred.

Therefore:

Autonomy should not automatically determine attribution.

33. Machine Errors

Consider an automated procurement system.

The company programs:

“Purchase 100 units if price ≤ AED 100.”

A software error causes:

“Purchase 10,000 units.”

Who is responsible?

Possible legal questions include:

• Was the system correctly programmed?
• Was the company negligent?
• Was the supplier aware of the error?
• Was the transaction electronically authenticated?
• Was there a contractual mechanism for correction?
• Was the error immediately communicated?
• Did the counterparty rely on the transaction?

The machine's “decision” does not itself answer these questions.

34. Autonomous Systems and Causation

Machine agency also affects causation.

Suppose:

AI recommends a transaction → employee approves → loss occurs.

Or:

AI executes transaction automatically → loss occurs.

The legal analysis differs because the degree of human intervention differs.

A court may need to distinguish:

cause of event

from

legal responsibility for event.

35. Legal Subjectivity and Digital Assets

Digital assets make the debate more complicated.

A machine might:

• control a crypto wallet;
• automatically trade tokens;
• manage a DAO treasury;
• execute smart contracts.

But control does not necessarily equal ownership.

The Techteryx proceedings demonstrate the importance of identifying the legally recognised beneficial/proprietary interests underlying digital assets. The court dealt with claims concerning the beneficial ownership of approximately USD 456 million and granted proprietary and freezing relief.

Thus:

Technical control ≠ legal ownership.

36. DAO and Machine Subjectivity

A DAO or autonomous protocol creates a particularly difficult problem.

It may appear that:

“The code operates itself.”

But civil law must identify:

• developers;
• governance participants;
• token holders;
• operators;
• legal entity;
• contractual relationships;
• property interests.

The mere existence of autonomous code does not automatically establish legal personality.

37. Arguments Supporting Machine Legal Personality

The theoretical arguments include:

A. Functional autonomy

Advanced systems may make decisions without immediate human intervention.

B. Economic autonomy

Some systems may control substantial digital assets.

C. Responsibility allocation

Personhood could theoretically create a clear liability structure.

D. Technological neutrality

If legal personality is functional rather than biological, sophisticated machines could theoretically qualify.

E. Commercial convenience

Separate machine entities could potentially contract continuously without constant human intervention.

These are academic and policy arguments, not a statement of current UAE law.

38. Arguments Against Machine Legal Personality

A. Lack of genuine legal intention

Machines do not necessarily possess human-style intention.

B. Attribution remains possible without personhood

Existing agency and corporate mechanisms can allocate responsibility.

C. Patrimony problem

A legal person needs an economically meaningful asset base.

D. Accountability problem

Developers and operators could attempt to hide behind the machine.

E. Regulatory complexity

Machine personhood would raise questions about:

• tax;
• licensing;
• insolvency;
• insurance;
• AML;
• sanctions;
• consumer protection;
• employment;
• data protection.

F. No present UAE statutory recognition

The current Civil Transactions Law specifies legally recognised categories of legal persons without creating a general category of AI/machine legal persons.

39. The UAE's Current Functional Approach

The available statutory and judicial framework supports a functional rather than personality-based approach.

In simplified form:

Stage 1

Machine performs an act.

Stage 2

Law determines whether the act creates legal consequences.

Stage 3

Law identifies the relevant human or legal person.

Stage 4

Contract/tort/property law determines rights and obligations.

Stage 5

Court determines disputes.

This avoids automatically converting technological autonomy into legal personality.

40. Important Distinction: Legal Person vs Legal Instrument

Article 92 of the current Civil Transactions Law is particularly important because it establishes the statutory basis for legal personality rather than simply treating every autonomous entity as a person.

41. Six Core Legal Questions for Any Machine-Agent Dispute

When an AI or automated system is involved, ask:

Question 1 — Who owns the system?

Developer, company, individual or other entity?

Question 2 — Who controls the system?

Technical ownership and operational control may differ.

Question 3 — Who authorised the system?

Was it authorised to conclude transactions?

Question 4 — What did the system actually do?

This is a factual/evidentiary question.

Question 5 — Who benefited?

Economic benefit may help identify the relevant legal relationship, although it is not alone determinative.

Question 6 — Who bears legal responsibility?

That depends on the applicable contract, legislation, agency principles, tort rules and other relevant law.

42. Practical Example

Facts

Company A develops an AI procurement system.

Company A programs it to purchase:

100 units whenever price falls below AED 50.

The AI mistakenly purchases:

100,000 units.

Possible arguments

Company A:

“The AI made the mistake.”

Supplier:

“The electronic system made a valid order.”

AI developer:

“We only supplied software.”

Legal analysis

The court could investigate:

1. programming instructions;
2. authority;
3. electronic authentication;
4. contractual terms;
5. system logs;
6. communications;
7. error detection;
8. notification;
9. reliance;
10. applicable remedies.

The AI's autonomy does not itself resolve the dispute.

43. Practical Example: Autonomous Trading

An AI trading system automatically purchases securities.

The system generates a loss of AED 20 million.

The relevant legal questions may include:

• Who authorised trading?
• What mandate existed?
• What regulatory rules applied?
• Was the system within its programmed limits?
• Did the operator supervise it?
• Was there a software defect?
• Was the transaction properly authenticated?
• Who owned the trading account?

Again:

The AI's autonomous operation is a factual circumstance, not automatically a legal personality rule.

44. Practical Example: Autonomous Vehicle

An autonomous vehicle causes injury.

Possible responsible persons/entities:

• vehicle owner;
• manufacturer;
• software developer;
• maintenance provider;
• data provider;
• operator;
• insurer.

The law must determine causation and responsibility.

Creating a machine legal person would not necessarily remove the need to examine these human and corporate relationships.

45. Future UAE Legislative Questions

Future legislation could potentially address:

1. Registration of autonomous systems

Should certain high-risk AI systems be registered?

2. Mandatory insurance

Should autonomous systems have compulsory liability insurance?

3. Dedicated asset pools

Should an autonomous system operate with separately allocated capital?

4. Human accountability

Should every high-risk AI system have an identifiable responsible legal person?

5. Auditability

Should system logs be preserved?

6. Explainability

Should certain automated decisions be explainable?

7. Digital-agent contracts

Should the law establish explicit rules for AI negotiation and contracting?

8. Limited machine personality

Should some autonomous systems receive narrowly defined legal capacity?

These are future-law questions, not current UAE legal rules.

46. Relationship Between Machine Agency and Legal Subjectivity

The debate can be represented as:

Automation

↓

Machine agency

↓

Legal attribution

↓

Liability

↓

Legal personality?

The last step does not automatically follow from the earlier steps.

A machine can be:

highly autonomous → legally consequential → economically valuable

without becoming:

a legal person.

47. Case-Law Synthesis

48. Mainland UAE and DIFC/ADGM Must Be Distinguished

This distinction is essential for examination and research.

Mainland UAE

The current civil-law framework is the Federal Decree by Law No. 25 of 2025, effective from 1 June 2026.

Its legal-person provisions expressly identify recognised categories of legal persons.

Electronic automated transactions are separately recognised under Federal Decree-Law No. 46 of 2021.

DIFC

The DIFC has its own legal framework and specialist Digital Economy Court. Part 58 expressly covers AI, blockchain, digital assets and automated dispute-resolution processes.

Important qualification

DIFC decisions are not automatically binding precedents for mainland UAE courts.

They are particularly useful for understanding how a UAE-based specialist common-law jurisdiction approaches emerging technology.

49. Overall Legal Position

The present UAE position can be summarised in five propositions:

Proposition 1

Automated transactions are legally recognised.

Federal Decree-Law No. 46 of 2021 expressly recognises automated electronic contracting.

Proposition 2

Automation does not itself create legal personality.

Proposition 3

Current UAE civil-law legal-person provisions identify legally recognised categories rather than AI/machines as a general category.

Proposition 4

Courts can impose legal consequences upon humans and companies for conduct involving automated systems.

The technology does not automatically become the bearer of liability.

Proposition 5

The DIFC has created a sophisticated procedural environment for technology disputes without equating technological autonomy with legal subjectivity.

Its Digital Economy Court expressly handles AI and other emerging-technology disputes.

50. Exam Revision Summary

Meaning

Machine agency = ability of an automated system to perform functions that have legal consequences.

Legal subjectivity = recognition of the machine itself as a holder of rights and duties.

Current UAE position

• Automated electronic transactions are recognised.
• AI and machines are not generally recognised as independent legal persons.
• Legal consequences can be attributed to human or corporate actors.
• Legal personality requires a legal basis.
• Digital assets can be legally protected without making the technology itself a legal person.
• AI-assisted professional conduct remains subject to human/legal accountability.
• DIFC has specialised Digital Economy Court jurisdiction for AI and other digital disputes.

Key formula

Machine Autonomy ≠ Legal Personality

and:

Automated Action + Legal Attribution = Legal Consequence

but:

Automated Action ≠ Independent Legal Subject

Conclusion

The UAE's emerging civil-law framework currently adopts a distinction between technological agency and legal subjectivity. Federal law permits automated electronic systems to participate in legally effective transactions, while the current Civil Transactions Law continues to define legal personality through legally recognised categories.

The major legal debate is therefore not simply whether machines can act autonomously, but who bears the legal consequences of that autonomous action.

For present UAE law, the more defensible framework is:

The machine may perform the act; the law identifies the legally responsible person or entity.

A future move toward machine legal personality would require much more than recognising autonomous decision-making. It would require rules concerning capacity, patrimony, ownership, liability, representation, insurance, insolvency, taxation, regulatory responsibility and access to courts. The present UAE framework has not generally made that transition.