Civil Law and UAE Legal Personality Distributed Across Digital Agents

1. Introduction

Legal personality distributed across digital agents is an emerging civil-law question concerning situations in which legally relevant activity is performed not by a single human actor, but through a combination of:

• human users;
• companies;
• artificial-intelligence systems;
• software agents;
• smart contracts;
• autonomous digital systems;
• blockchain networks;
• digital wallets;
• platforms; and
• automated decision-making systems.

The central legal question is:

When a digital agent performs an action, who is the legally responsible person?

Under the present UAE legal framework, the more appropriate approach is generally not to treat AI or software itself as automatically possessing independent legal personality. Instead, legal personality and responsibility ordinarily remain attached to recognised natural or juridical persons, while the conduct of a digital system is attributed through applicable rules concerning agency, contracts, electronic transactions, corporate personality, authority, ownership, control, fault and liability.

This subject is especially relevant to the UAE because the DIFC Digital Economy Court expressly deals with disputes involving AI, blockchain, digital assets, decentralised autonomous organisations (DAOs), decentralised finance (DeFi), decentralised applications (DApps), smart contracts, digital signatures and digital identification.

2. Meaning of Distributed Legal Personality

Traditional civil law generally begins with two principal categories:

Natural person

A human being possessing legal capacity.

Juridical person

An entity recognised by law as having a separate legal personality, such as:

• a company;
• certain foundations;
• other legally recognised entities.

Digital agents introduce a third technological layer:

Digital agent

Software capable of performing actions or making decisions according to programmed rules, data, instructions or AI models.

Examples include:

• AI trading agents;
• automated contract systems;
• smart-contract systems;
• autonomous customer-service agents;
• blockchain bots;
• algorithmic investment systems.

The important distinction is:

Technological autonomy does not necessarily equal legal personality.

3. Legal Personality Versus Technological Agency

These concepts should not be confused.

Legal personality

Means that the law recognises an entity as capable of holding rights and obligations.

Technological agency

Means that software can perform an action with some degree of autonomy.

An AI system may:

• send an email;
• conclude an automated transaction;
• generate a document;
• execute code;
• transfer a digital asset.

But that does not automatically mean that the AI itself:

• owns property;
• owes a debt;
• possesses contractual capacity;
• can sue;
• can be sued;
• bears civil liability.

The legal system must determine which legally recognised person stands behind the technological activity.

4. UAE Civil-Law Starting Point

The UAE civil-law framework recognises legal personality primarily through human persons and legally recognised juridical entities.

Consequently, a digital agent generally operates as a technological instrument or autonomous mechanism connected to a legal person, rather than automatically becoming a new juridical person.

For example:

Company A deploys an AI procurement agent.

The AI agent automatically purchases goods.

If the transaction is legally binding, the initial question is not necessarily:

"Is the AI itself a legal person?"

Instead:

"What legal relationship exists between Company A, the supplier and the AI system, and was the AI authorised or configured to act on Company A's behalf?"

5. Distributed Agency Model

A useful conceptual model is:

Human → Company → Digital Platform → AI Agent → Smart Contract → Blockchain

Responsibility may therefore be distributed across several legal and technological layers.

For example:

1. shareholders own Company A;
2. Company A employs a technology provider;
3. the provider creates an AI agent;
4. the AI agent interacts with a platform;
5. the platform activates a smart contract;
6. the smart contract transfers digital assets.

If something goes wrong, the court may need to determine:

• who controlled the system;
• who authorised the transaction;
• who benefited;
• who supplied the software;
• whether the system malfunctioned;
• whether instructions were defective;
• whether a contract existed;
• whether statutory duties were breached.

6. AI Does Not Automatically Become a Legal Person

A critical distinction is necessary.

Human-controlled AI

A person uses AI as a tool.

Corporate AI

A company deploys AI in its business.

Autonomous AI

The system can perform actions without immediate human intervention.

Distributed AI

Multiple systems collectively contribute to an outcome.

None of these characteristics necessarily creates legal personality.

The UAE legal analysis should therefore begin with attribution, rather than immediately creating a new category of legal person.

7. Attribution of Digital Actions

Attribution means determining whose legally relevant conduct an automated action represents.

A court may consider:

• ownership;
• control;
• authority;
• contractual terms;
• system architecture;
• instructions;
• access credentials;
• authentication;
• economic benefit;
• foreseeable operation;
• cybersecurity controls.

Example

A company gives an AI agent authority to purchase raw materials up to AED 500,000.

The agent purchases AED 400,000 of materials.

The fact that the AI made the selection autonomously does not necessarily prevent the transaction from being attributable to the company.

The key question becomes:

Was the system acting within the authority granted to it?

8. Electronic Transactions

UAE electronic-transactions legislation is particularly relevant.

Electronic systems can perform legally significant actions without a person manually typing every part of the transaction.

The legal significance of an electronic communication or signature therefore depends upon:

• authentication;
• intent;
• attribution;
• reliability;
• statutory requirements.

The DIFC courts have developed useful jurisprudence concerning electronic signatures and electronic communications.

9. Case Law 1 — Ondina v Olin, CFI 046/2025

In Ondina v Olin, the DIFC Court considered whether email correspondence could satisfy a statutory requirement for a written and signed amendment.

The Court considered the DIFC Electronic Transactions Law and concluded that the relevant electronic communication could constitute an electronic signature where the circumstances demonstrated the required intention. (difccourts.ae)

Relevance to digital agents

The case demonstrates an important principle:

The law may focus on attribution and intention, rather than requiring a traditional handwritten signature.

For autonomous digital agents, this raises the corresponding question:

Whose intention and authority does the digital action represent?

10. Case Law 2 — ICICI Bank Ltd v Bavaguthu Raghuram Shetty, CFI 034/2022

In ICICI Bank Ltd v Bavaguthu Raghuram Shetty, the DIFC Court considered disputed electronic signatures and expert evidence regarding their authenticity and authorisation.

The Court's analysis illustrates that an electronically applied signature cannot simply be characterised as fraudulent merely because it is digital. The question includes whether it was authorised and attributable to the relevant person. (difccourts.ae)

Relevance

For digital agents, attribution can similarly depend upon:

• who controlled the authentication mechanism;
• who authorised the transaction;
• how the system was configured;
• whether the transaction was within authority.

11. Case Law 3 — Naho v Neukirchi, DIFC SCT 415/2024

In Naho v Neukirchi, the DIFC Small Claims Tribunal considered the legal significance of electronic communications and electronic signatures under the DIFC Electronic Transactions Law.

The case illustrates that electronic conduct may have legally binding consequences where the statutory requirements for electronic transactions are satisfied. (difccourts.ae)

Relevance

A digital agent can potentially become part of a legally significant electronic transaction without itself becoming a separate legal person.

The legal question remains:

Which person is legally connected to the system's action?

12. Case Law 4 — Arabyads Holding Limited v Gulrez Alam Marghoob Alam, [2025] ADGMCFI 0032

This case is especially important for AI.

The ADGM Court dealt with legal submissions containing authorities that were identified as non-existent and associated with AI-assisted legal research. The court imposed a substantial wasted-costs consequence against the relevant law firm. (jibudocs.com)

Relevance

The case demonstrates:

AI activity does not transfer professional responsibility from the human legal practitioner to the AI system.

The human lawyer remained responsible for verifying the material.

This provides a strong analogy for distributed legal agency:

AI performs → human/legal entity remains accountable.

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

Techteryx Ltd v Aria Commodities DMCC & Others is an important Digital Economy Court authority.

The proceedings involved complex commercial and digital-economy issues and were dealt with within the DIFC's specialist Digital Economy Court framework. (difccourts.ae)

Relevance

The importance of this authority lies in the institutional recognition that digital-economy disputes can require specialised judicial treatment.

The Digital Economy Court's jurisdiction includes disputes involving:

• AI;
• blockchain;
• smart contracts;
• digital assets;
• DAOs;
• DeFi;
• DApps.

This provides a judicial environment in which questions concerning distributed digital agency can be addressed.

14. Case Law 6 — World Group Holding Ltd v Sajid Barkat Al Barkat, [2021] DIFC CFI 087

In World Group Holding Ltd v Sajid Barkat Al Barkat, witnesses gave evidence remotely by video link.

The Court evaluated the evidence using ordinary judicial principles, including examination and credibility assessment. (difccourts.ae)

Relevance

This case demonstrates that technology can change the mode of participation without necessarily changing the underlying legal personality of the participants.

Likewise, an AI interface may change how a person acts without necessarily becoming the legal person itself.

15. Additional Case — Union Bank of India v Velocity Industries LLC, [2020] DIFC CFI 025

In Union Bank of India (DIFC Branch) v Velocity Industries LLC, the DIFC Court dealt with remote witness evidence from India.

The Court addressed technological participation while preserving the fundamental procedural role of the court in determining evidentiary questions. (difccourts.ae)

Relevance

Technology can distribute participation, but the law still requires identifiable legal actors.

16. Digital Agents and Contract Formation

Consider:

Company A deploys an AI purchasing agent.

The agent:

1. searches suppliers;
2. compares prices;
3. selects Supplier B;
4. accepts Supplier B's electronic offer;
5. generates an electronic purchase order.

Was there a contract?

The legal analysis should consider:

Step 1 — Authority

Did Company A authorise the system?

Step 2 — Offer

Was there a legally valid offer?

Step 3 — Acceptance

Did the automated action constitute acceptance?

Step 4 — Attribution

Can the acceptance be attributed to Company A?

Step 5 — Applicable electronic-transactions law

Were the statutory requirements satisfied?

The analysis does not necessarily require the AI to possess independent legal personality.

17. Digital Agents as Automated Representatives

One possible legal model is to treat the digital agent as a technological extension of a principal.

Principal

Company A

↓

Digital agent

AI purchasing system

↓

Third party

Supplier B

The legal relationship remains:

Company A ↔ Supplier B

The AI is the mechanism through which the relationship operates.

This is conceptually similar to other forms of automation, although highly autonomous AI systems introduce additional questions concerning foreseeability and control.

18. When the Agent Exceeds Its Authority

Suppose Company A authorises an AI agent to purchase goods up to AED 1 million.

The AI purchases AED 10 million.

Questions arise:

• Was the transaction binding?
• Did the supplier know of the limitation?
• Was the system's authority externally communicated?
• Was the AI compromised?
• Did Company A negligently configure the system?
• Did the supplier reasonably rely upon the transaction?

This illustrates why digital agency and authority will become increasingly important areas of UAE civil law.

19. AI Errors and Civil Liability

AI systems can produce:

• incorrect recommendations;
• inaccurate classifications;
• defective decisions;
• erroneous transactions;
• misleading information.

Civil liability may potentially involve several parties:

AI developer

Did the software contain a defect?

Platform operator

Did the platform fail to implement reasonable controls?

User

Was the system negligently used?

Company

Did it fail to supervise the system?

Third party

Did it rely unreasonably upon an automated output?

The correct allocation depends on the applicable contract, statutory provisions, evidence and circumstances.

20. Distributed Liability

This produces a concept that can be called:

Distributed civil responsibility

Example:

Developer → Platform → Company → AI Agent → Customer

Each layer may potentially have a different legal relationship.

The court should therefore avoid the simplistic assumption:

"The AI made the decision, so nobody is liable."

Instead, the court should identify the relevant legal relationships and applicable rules.

21. Autonomous Digital Agents and Companies

A company is already a recognised juridical person.

Therefore, a company can:

• own property;
• enter contracts;
• incur debts;
• sue;
• be sued.

The company's use of autonomous technology does not necessarily change its juridical personality.

For example:

Company A uses an autonomous AI system to manage its logistics.

The AI does not automatically replace Company A as the legal person.

22. DAOs and Distributed Legal Personality

The question becomes more difficult with Decentralised Autonomous Organisations (DAOs).

A DAO may involve:

• blockchain governance;
• token holders;
• smart contracts;
• decentralised voting;
• automated treasury management.

Who is the legal person?

Possibilities may include:

1. an incorporated entity associated with the DAO;
2. individual participants;
3. contractual relationships between participants;
4. trustees or managers;
5. persons controlling relevant systems;
6. another legally recognised structure.

A blockchain protocol's existence alone does not necessarily create an independent juridical person.

23. Smart Contracts

A smart contract is generally software that automatically executes specified functions when predetermined conditions occur.

For example:

If payment is received, release the digital asset.

The code performs the execution.

But legal questions remain:

• Who are the parties?
• What legal agreement does the code represent?
• What happens if the code contains an error?
• Can the transaction be reversed?
• Who bears the loss?
• What law applies?
• Which court has jurisdiction?

Thus:

Code execution ≠ complete legal analysis.

24. AI Agents and Corporate Authority

Traditional corporate authority may involve:

• directors;
• managers;
• authorised signatories;
• powers of attorney.

Digital agents introduce another layer:

Machine authority

The machine may have technical permission to act.

But technical permission does not necessarily equal legal authority.

For example:

API credential → technical authority

does not automatically establish:

legal authority → power to bind a company.

The contract, corporate documents, applicable law and surrounding circumstances remain relevant.

25. Digital Identity

Distributed digital agency also depends upon identity.

A legal system may need to determine:

• who controls an account;
• who owns a digital wallet;
• who authorised a transaction;
• who controlled a private key;
• who operated an AI agent;
• whether an identity was compromised.

Digital identity therefore acts as a bridge between:

technology → legal person.

26. Cybersecurity and Attribution

Suppose an AI agent transfers AED 2 million because an attacker compromised the system.

Potential questions include:

• Who authorised the original system?
• Was authentication compromised?
• Was the company negligent?
• Was the attack foreseeable?
• Did the third party know of the compromise?
• Was there an obligation to maintain cybersecurity controls?

The distinction between:

authorised autonomous action

and

unauthorised automated action

can become decisive.

27. Digital Agents and Evidence

When a digital agent acts, the court may need evidence concerning:

• system logs;
• source code;
• API records;
• authentication data;
• access logs;
• timestamps;
• configuration settings;
• blockchain transactions;
• server records;
• communications.

The UAE's electronic-evidence framework is therefore central to future disputes involving autonomous systems.

28. Legal Personality and Accountability

Granting legal personality to an AI system would potentially separate:

AI system → liability

from:

developer/user/company → liability

But this raises difficult questions.

If an AI has no:

• independent assets;
• bank account;
• insurance;
• human governance;
• ability to pay damages,

then giving it legal personality may not necessarily solve the practical problem of compensation.

The more immediate civil-law issue is often:

Who should bear responsibility for deploying, controlling or benefiting from the system?

29. Should AI Receive Separate Legal Personality?

This is a major theoretical debate.

Arguments sometimes advanced for separate personality

• AI can operate autonomously.
• AI can enter numerous transactions.
• AI systems may become increasingly independent.
• Separate personality could theoretically simplify attribution.

Arguments against immediate recognition

• AI lacks conventional human legal capacity.
• Responsibility can usually be allocated to existing persons.
• AI may not possess independent assets.
• Legal personality could obscure accountability.
• Developers and users might attempt to use AI personality to avoid liability.

Under the present UAE framework, there is no general rule automatically granting ordinary AI systems independent juridical personality.

30. Human-in-the-Loop Model

A practical UAE model is:

AI → recommendation/action → human supervision → legally attributable decision

This model is particularly consistent with current UAE judicial AI developments.

The Abu Dhabi Judicial Department's AI Judicial Platform initiative has been described as operating with human supervision and verification, rather than replacing human judicial responsibility.

This reinforces the principle that technological autonomy does not automatically eliminate human accountability.

31. Human-on-the-Loop Model

A more advanced model involves:

AI operates autonomously → humans monitor → humans intervene when necessary

This is particularly relevant to:

• financial trading;
• logistics;
• smart buildings;
• automated procurement;
• cybersecurity.

The legal question becomes whether the human organisation implemented adequate:

• controls;
• monitoring;
• intervention mechanisms;
• risk management.

32. Fully Autonomous Digital Systems

The most difficult case is:

AI Agent → AI Agent → Smart Contract → Blockchain

with no immediate human intervention.

For example:

1. Agent A identifies a commercial opportunity.
2. Agent A contacts Agent B.
3. Agent B negotiates automatically.
4. Smart contract executes.
5. Digital assets transfer.

A future UAE court may need to determine:

• who programmed the agents;
• who deployed them;
• what authority was granted;
• who owned the assets;
• who benefited;
• whether the agents' interaction constituted a contract;
• whether any party acted negligently.

Again, the likely legal focus is attribution and liability, not merely technological autonomy.

33. Civil-Law Principles Applicable to Digital Agents

Several traditional civil-law principles remain relevant.

Good faith

Automated execution should not automatically defeat good-faith contractual obligations.

Consent

The system's operation must be connected to legally recognisable consent or authority.

Causation

The court must determine whether the conduct caused the loss.

Fault

Where liability depends on fault, system design and supervision may become relevant.

Compensation

The injured party may seek legally available remedies.

Contractual interpretation

The court may need to interpret the human agreement underlying automated code.

34. Distributed Legal Personality in the UAE — Analytical Model

A useful framework is:

Layer 1 — Human persons

Founders, shareholders, directors, employees and users.

Layer 2 — Juridical persons

Companies and other recognised entities.

Layer 3 — Digital systems

Platforms, AI agents and smart contracts.

Layer 4 — Distributed infrastructure

Blockchain networks, decentralised protocols and digital assets.

Layer 5 — Legal attribution

The law connects the technological action to one or more recognised legal actors.

Thus:

Technology may distribute activity without necessarily distributing legal personality.

35. Practical Example

Imagine a UAE company operates an AI import agent.

The system:

1. identifies a supplier;
2. negotiates automatically;
3. signs an electronic purchase order;
4. transfers cryptocurrency;
5. receives goods;
6. discovers that the goods are defective.

The civil-law analysis might proceed:

Question 1

Who owns the AI system?

Question 2

Who authorised its operation?

Question 3

Was the supplier aware it was dealing with an automated system?

Question 4

Was the electronic contract valid?

Question 5

Was the crypto payment attributable to the company?

Question 6

Who bears the loss from defective goods?

Question 7

Was the AI acting within its programmed authority?

Question 8

Was there negligence in system design or supervision?

This demonstrates why distributed digital agency creates a multi-layer attribution problem.

36. Importance for UAE Lawyers

Civil practitioners increasingly need competence in:

• electronic transactions;
• AI;
• blockchain;
• smart contracts;
• digital evidence;
• cybersecurity;
• data protection;
• digital assets;
• electronic signatures;
• algorithmic decision-making.

A lawyer handling a digital-agent dispute cannot rely exclusively on traditional contract analysis.

They may need to understand:

legal rule + technical architecture + evidence + attribution.

37. Importance for Contract Drafting

Contracts involving AI agents should consider:

Authority

What can the agent do?

Spending limits

What monetary limits apply?

Human approval

Which transactions require human confirmation?

Auditability

Must system actions be logged?

Cybersecurity

Who maintains security?

Errors

Who bears losses from algorithmic errors?

Suspension

When can the system be disabled?

Data

Who owns and controls data?

Termination

What happens to the agent after termination?

Liability

How is responsibility allocated?

38. Future UAE Legal Development

Potential future developments may include:

1. legal rules for autonomous agents;
2. specific AI-liability legislation;
3. regulatory frameworks for agentic AI;
4. AI identity standards;
5. machine-readable contracts;
6. smart-contract dispute mechanisms;
7. digital-agent licensing;
8. rules for DAO legal status;
9. automated enforcement mechanisms;
10. specialised AI courts or judicial divisions.

The UAE's existing Digital Economy Court and AI judicial initiatives provide institutional foundations for these developments.

39. Six-Case Revision Table

40. Examination Points

For examination purposes, remember:

Legal personality

A legal status recognised by law allowing an entity to possess rights and obligations.

Digital agent

A software system capable of performing actions with varying degrees of autonomy.

Core issue

Attribution of automated conduct to a legally recognised person.

Main principles

1. AI autonomy does not automatically create legal personality.
2. Legal responsibility can remain with the developer, owner, user or company.
3. Corporate personality remains distinct from the technological systems used by a company.
4. Electronic signatures can have legal significance.
5. Digital transactions require authentication and attribution.
6. Smart-contract execution does not necessarily answer all legal questions.
7. DAO structures create difficult questions concerning legal status.
8. Digital evidence is essential to determining attribution.
9. AI does not eliminate professional responsibility.
10. Human oversight remains important in judicial AI.

41. Conclusion

The concept of legal personality distributed across digital agents represents a major challenge for contemporary UAE civil law.

The fundamental distinction is:

Digital autonomy is not the same thing as legal personality.

An AI system may act autonomously, negotiate, generate documents, execute transactions or interact with another automated system. Nevertheless, the law must still determine which recognised legal person owns, controls, authorises, benefits from or is otherwise legally connected to that activity.

The emerging UAE framework points toward a model of distributed technological activity but legally attributable responsibility. The DIFC Digital Economy Court, electronic-transactions jurisprudence, AI-related litigation such as Arabyads, and the UAE's development of AI-enabled judicial systems demonstrate that the issue is already moving from legal theory into practical civil justice.

For civil-law purposes, the most important principle is therefore:

The increasing autonomy of digital agents does not by itself require the immediate creation of a new category of legal person; the central task is to develop reliable rules for attribution, authority, evidence, liability and accountability when legally significant activity is performed through autonomous digital systems.