Civil Law and UAE Non-Human Intentionality in Legal Analysis

1. Introduction

Non-human intentionality refers to situations where conduct appears to be generated by something other than a natural human decision-maker—for example:

artificial intelligence (AI);

automated software;

smart contracts;

algorithms;

bots and autonomous systems;

corporate entities acting through organizational processes;

animals or other non-human agents;

machine-generated communications or transactions.

The central legal question is:

Can a non-human system itself possess legally relevant intention, knowledge, consent, negligence, or responsibility under UAE civil law?

Under UAE civil-law principles, the safer legal position is that machines, algorithms and AI systems are not presently treated as independent legal persons possessing human-style legal intention. Instead, the law normally looks through the technology to identify the human or legal person who owns, controls, deploys, authorizes, benefits from, or is otherwise legally responsible for the system.

Thus:

Machine action ≠ automatically machine intention.

Instead:

Machine action → attribution → human/legal person → legal intention/duty → liability.

2. Meaning of Non-Human Intentionality

Definition

Non-human intentionality may be understood as:

The apparent capacity of a non-human system or entity to select, initiate, communicate, or execute conduct that appears purposeful, without a contemporaneous direct human instruction for each individual act.

For example, an AI trading system may automatically place an order. A chatbot may make a contractual representation. A smart contract may automatically transfer digital assets when programmed conditions are satisfied.

The system appears to "decide."

But legally several different questions must be separated:

Who created the system?

Who deployed it?

Who controlled it?

Who authorized its operation?

Who benefited from the transaction?

Was the system operating as programmed?

Was there human intervention?

Was there a legal duty concerning the system?

Who suffered the resulting loss?

Can the resulting conduct legally be attributed to a person or company?

3. Fundamental UAE Civil-Law Principle

Traditional civil law is built around legally recognized persons.

These include:

natural persons; and

juridical/legal persons such as companies.

An AI system, algorithm or autonomous software program does not automatically become a separate juridical person merely because it can perform complex operations.

Therefore:

Traditional model

Human intention → legal act → legal consequence

Digital model

Human/legal-person design → algorithm → automated conduct → legal attribution → legal consequence

This distinction is extremely important.

4. Intention and Legal Personality Are Different

A system may exhibit what appears to be purposeful behaviour without possessing legal personality.

For example:

An AI system is programmed to purchase goods whenever the price falls below AED 100.

The system purchases goods at AED 90.

Technically, the system selected the action.

Legally, however, the questions include:

Who owns the account?

Who programmed the system?

Who authorized the purchase?

Was the account used by an employee?

Was the AI operating within its authorized parameters?

Was the transaction electronically authenticated?

Was there fraud?

Was the system compromised?

The legal analysis therefore concerns attribution, rather than treating the AI as an independent human mind.

5. Non-Human Intentionality and UAE Civil Law

The concept intersects with several areas of UAE law:

6. Non-Human Intentionality Is Not the Same as Legal Intention

This distinction should be remembered for examinations.

Human psychological intention

A human may:

know;

intend;

believe;

understand;

deceive;

consent;

make a mistake.

Machine functionality

A machine may:

calculate;

classify;

predict;

generate;

execute;

recommend;

trigger;

communicate.

A machine can therefore simulate or produce intentional-looking behaviour without possessing legally recognized subjective intention.

Formula

Apparent Machine Intention ≠ Legal Intention

Instead:

Machine Conduct + Attribution + Applicable Legal Rule = Legal Consequence

7. Electronic Transactions and Automated Systems

UAE electronic-transactions legislation provides an important foundation for dealing with electronically generated communications and transactions.

The legal system increasingly recognizes:

electronic records;

electronic signatures;

electronic communications;

automated electronic processes;

authentication mechanisms;

trust services.

The important issue is not whether a computer "has a mind."

The issue is whether the electronic act can legally be attributed to a person or organization.

8. Automated Contract Formation

Suppose Company A operates an online platform.

A customer clicks "Buy."

The system automatically:

receives the order;

verifies payment;

accepts the order;

generates an invoice;

sends confirmation.

There may be no employee personally reviewing the transaction.

Nevertheless, the legal transaction may still be attributable to the company.

Legal structure

Company authorization

↓

Automated system

↓

Electronic communication

↓

Customer acceptance/payment

↓

Contractual consequences

The software does not necessarily become the contracting party.

9. AI-Generated Statements

Suppose an AI chatbot operated by a company tells a customer:

"This product is guaranteed for five years."

The customer relies upon the statement and purchases the product.

The AI subsequently generates an incorrect statement.

The legal analysis should ask:

Was the chatbot authorized by the company?

Was it presented as the company's representative?

Was the statement reasonably relied upon?

Was the information misleading?

Was there a contractual representation?

Was there negligence in deploying the system?

Did the company know about recurring errors?

Was adequate human supervision provided?

The AI's "intention" is normally less important than the legal responsibility of the organization behind the system.

10. AI and Agency

Agency provides an especially useful conceptual framework.

Traditional agency:

Principal → Agent → Third Party

Digital environment:

Principal → Software/AI System → Third Party

The difficult question is whether the software itself should be regarded as an agent.

A more conservative UAE civil-law approach is:

The software functions as an instrument or automated mechanism through which the principal's authorized conduct is expressed, rather than automatically becoming an independent juridical agent.

Therefore, the legal analysis focuses on:

authority;

authorization;

attribution;

apparent authority;

system configuration;

human supervision;

scope of mandate.

11. Corporate Persons and Non-Human Intention

A company itself is also not a natural human being.

Yet a company possesses separate legal personality.

This demonstrates an important distinction:

"Non-human" does not necessarily mean "without legal personality."

A corporation is legally recognized because legislation gives it juridical personality.

An AI system generally does not receive such independent legal personality merely because it operates autonomously.

Therefore:

Company

Non-human + legally recognized juridical person = legal personality

AI

Non-human + no independent juridical personality merely by operation = generally no independent civil liability

12. AI Liability Through Human or Corporate Attribution

A useful model is:

AI Liability Chain

Design

↓

Deployment

↓

Authorization

↓

Operation

↓

Harm

↓

Attribution

↓

Responsible Person/Entity

↓

Civil Remedy

This prevents the legal system from simply saying:

"The AI did it."

The law must identify the legally responsible actor.

13. Non-Human Intentionality and Causation

Automated systems create difficult causation questions.

Consider:

Developer creates an algorithm.

Company deploys it.

Cloud provider hosts it.

Data provider supplies information.

AI makes a recommendation.

Employee relies on the recommendation.

Customer suffers damage.

Who caused the damage?

The analysis requires:

duty;

breach;

factual causation;

legal causation/remoteness;

damage;

attribution;

possible intervening causes.

Formula

AI-related Civil Liability = Duty + Breach/Fault + Causation + Damage + Attribution

The mere fact that an algorithm was involved does not automatically establish liability.

14. Non-Human Intentionality and Mistake

AI systems can generate erroneous outcomes.

For example:

wrong price;

wrong identity;

wrong recommendation;

wrong classification;

incorrect automated acceptance.

This raises the question:

Does an algorithmic error amount to a legally relevant mistake?

Usually, the legal inquiry should focus on the human/legal person's consent and the statutory requirements governing mistake, rather than attributing a subjective mistake to the machine.

The system itself does not normally become the bearer of the contractual mistake.

15. Non-Human Intentionality and Fraud

Suppose an AI generates false information.

The AI does not automatically become a fraudulent legal actor.

The investigation should ask:

Who supplied the underlying information?

Who programmed the system?

Was the output knowingly manipulated?

Who knew of the false information?

Who authorized publication?

Did a person deliberately exploit the system?

Did the company fail to correct known defects?

Important distinction

Machine-generated falsehood ≠ automatically human fraud

But:

Human knowledge/manipulation + machine deployment + reliance + damage → potential civil liability

16. Non-Human Intentionality and Negligence

AI systems create a different problem where no human deliberately intended the harm.

Example:

A company deploys an automated credit-scoring system.

The system repeatedly produces inaccurate results because the underlying data are defective.

No employee intended to discriminate against or harm a customer.

Potential legal questions include:

Was there a duty of care?

Was the system adequately tested?

Was monitoring undertaken?

Was the data sufficiently accurate?

Was the defect foreseeable?

Was there a failure to supervise?

Did the failure cause actual damage?

This illustrates an important principle:

Civil liability does not always require subjective intention.

Depending on the applicable cause of action, negligence, breach of duty, contractual liability, statutory liability or another legal basis may be sufficient.

17. Non-Human Intentionality and Digital Evidence

Machine-generated evidence may include:

system logs;

timestamps;

metadata;

audit trails;

server records;

transaction histories;

electronic signatures;

blockchain records;

API logs;

automated emails;

AI-generated documents.

The important legal questions are:

Authenticity

Is the record genuine?

Attribution

Who or what generated it?

Integrity

Has it been altered?

Chronology

When was it created?

Reliability

How reliable is the underlying system?

Corroboration

Is other evidence consistent with it?

Formula

Digital Evidentiary Value = Authenticity + Attribution + Integrity + Chronology + Provenance + Corroboration

18. Case Law and Judicial Authorities

Because "non-human intentionality" is not a standalone named doctrine in UAE mainland civil law, there is limited direct UAE jurisprudence specifically declaring whether an AI has legal intention.

The following cases are therefore best understood as relevant authorities on electronic contracting, attribution, authority, evidence, consent and automated/digital conduct, rather than as cases holding that AI possesses or lacks legal personality.

Case 1: ICICI Bank Ltd v Bavaguthu Raghuram Shetty

[2022] DIFC CFI 034

This case is important for electronic contracting and the legal significance of electronically communicated agreements.

It demonstrates that courts can examine:

electronic communications;

signatures;

contractual formation;

authority;

the circumstances surrounding electronic execution.

Relevance

The case supports the proposition that the legal question surrounding digital transactions is generally attribution and contractual intention of the relevant legal persons, rather than whether the computer system itself possessed subjective intention.

Case 2: GFH Capital Ltd v David Lawrence Haigh

[2014] DIFC CFI 020

The case involved electronic communications and questions concerning contractual authority and communications.

Relevance

It illustrates the importance of determining:

who communicated;

whether the communication was authorized;

whether the person had authority;

what legal effect should be given to electronic communications.

For AI systems, the same analytical structure becomes:

Who authorized the system? → What authority did it have? → What communication did it generate? → To whom should that conduct be attributed?

Case 3: Ondina v Olin

[2025] DIFC CFI 046

This decision is relevant to electronically communicated contractual conduct and questions surrounding electronic execution.

Relevance

It illustrates that digital execution must be analyzed through recognized legal concepts such as:

consent;

authentication;

contractual intention;

attribution;

evidence.

An automated system does not automatically become the legal holder of contractual intention merely because it technically generated a communication.

Case 4: Naho v Neukirchi

[2024] DIFC SCT 415

The decision involved electronic communications and electronic-signature issues.

Relevance

It demonstrates the importance of proving the connection between:

electronic record → person → authorization → transaction

This is particularly important when AI or automated systems generate communications.

Case 5: Tarig Mohamed Abdelsalam Abdelrahman v Expresso Telecom Group Ltd

[2021] DIFC CFI 056

This case concerned electronic service and the legal significance of electronic communications.

Relevance

The case illustrates that electronic communication can have procedural/legal consequences when the law recognizes the communication mechanism and the connection between the communication and the relevant party.

For automated systems, this reinforces the importance of attribution and reliability.

Case 6: Jonathan Lau v Qashio Holding Company Ltd & Armin Moradi Tosarvandani

[2026] DIFC CFI 058

This is particularly relevant to modern digital evidence.

The case involved electronic records, including native email material, audit trails and electronic-signature-related evidence.

Relevance

It demonstrates why courts dealing with digitally generated evidence may need to examine:

metadata;

audit trails;

electronic records;

authenticity;

provenance;

system-generated information.

For AI disputes, this type of evidence can help establish who controlled or authorized the relevant digital system.

Case 7: Dimension B+ Ltd v Saleh Abdelkarim Hussain Abdelrahman Almaazmi

[2024] DIFC CFI 094

The case concerned contractual consent and the effect of a signed integrated agreement.

Relevance

It reinforces the importance of identifying legally relevant consent and recognized vitiating factors.

This is significant for automated contracting because a party cannot necessarily avoid a transaction simply by saying:

"The computer generated it."

The court must examine the legally relevant circumstances surrounding authorization, consent and any recognized defect.

Case 8: Khaled Salem Musabeh Humaid Al Mheiri v John Cameron

[2025] DIFC CA 008

This appellate decision concerned UAE-law contractual issues, including the distinction between fundamental mistake and other forms of contractual error.

Relevance

It is useful when analyzing AI-generated mistakes.

The relevant question is not whether the software "made a mistake" psychologically, but whether the legally relevant party's consent was affected by a recognized mistake under the applicable UAE-law principles.

19. Significance of the Cases

The authorities collectively demonstrate an important analytical progression:

20. Smart Contracts and Non-Human Intentionality

Smart contracts provide an especially interesting example.

Suppose:

If payment is received, the software automatically transfers a digital asset.

No person presses the transfer button.

The system executes automatically.

The legal analysis should distinguish:

Programming intention

What did the programmer design?

Contractual intention

What did the contracting parties legally agree to?

Automated execution

What did the software actually do?

Legal consequence

What does applicable law recognize?

Therefore:

Code execution ≠ necessarily complete proof of contractual intention.

Code can be evidence of agreed terms, but legal interpretation may still require examination of:

contract text;

circumstances;

authority;

mistake;

fraud;

public policy;

mandatory law;

performance;

applicable remedies.

21. Autonomous AI Agents

Future AI systems may:

negotiate prices;

select suppliers;

execute purchases;

enter into digital transactions;

manage accounts;

make investment decisions;

generate legal documents;

initiate claims;

communicate with counterparties.

This produces a difficult question:

If an AI independently negotiates and concludes a transaction, who is legally bound?

A UAE civil-law analysis should initially examine:

Principal identity

System ownership

Authorization

Scope of authority

Electronic authentication

Contract formation

Capacity

Mistake

Fraud

Evidence

Reliance

Causation

Damage

Applicable statutory rules

22. Non-Human Intentionality and Corporate Liability

A corporation may deploy AI through several departments:

Board

↓

Management

↓

Technology department

↓

AI developer

↓

AI system

↓

Customer

If damage occurs, responsibility cannot automatically be assigned to the AI.

The investigation may instead consider:

corporate authorization;

employee authority;

contractual obligations;

internal controls;

cybersecurity;

data governance;

system testing;

supervision;

outsourcing arrangements;

contractual allocation of risk.

23. Multi-Actor AI Liability

Modern AI systems may involve:

model developer;

data provider;

cloud provider;

software integrator;

platform operator;

deploying company;

employee;

end user;

payment provider;

cybersecurity provider.

Therefore:

Multi-Actor Formula

AI Liability = Duty + Breach/Fault + Causation + Damage + Attribution

Where several actors contribute:

Total Harm → Causal Contributions → Individual Legal Bases → Allocation of Responsibility

Ownership of the AI model alone should not automatically establish liability for every consequence.

24. Non-Human Intentionality and Animals

The concept can also be applied to animals, although this is a different legal problem.

An animal may intentionally move, attack or damage property in an ordinary behavioural sense.

But civil law does not generally treat the animal as a human legal decision-maker.

The legal question becomes:

Who owned the animal?

Who controlled it?

Was there a duty of supervision?

Was the harm foreseeable?

Did the owner breach a statutory or civil obligation?

Thus:

Animal intention → factual behaviour

but:

Owner/controller → possible legal responsibility

The same analytical distinction helps explain AI systems.

25. Non-Human Intentionality and Legal Responsibility

A useful distinction is:

Level 1 — Behaviour

What did the system do?

Level 2 — Causation

Did that conduct cause the damage?

Level 3 — Attribution

Which person/entity is legally connected to that conduct?

Level 4 — Fault/Duty

Did that person/entity breach a relevant obligation?

Level 5 — Remedy

What compensation, restitution, injunction or other remedy is available?

Therefore:

Behaviour ≠ Liability

and:

Autonomy ≠ Legal Personality

26. Human-in-the-Loop and Human-on-the-Loop

Human-in-the-loop

A person approves each important decision.

Example:

AI recommends loan → employee approves → loan issued.

Attribution may be easier because a human makes the final decision.

Human-on-the-loop

AI operates automatically while humans monitor the system.

Example:

AI continuously monitors transactions and automatically blocks suspicious transactions.

Legal questions become more complicated:

Was monitoring adequate?

Were warnings ignored?

Was the system defective?

Was intervention reasonably possible?

Human-out-of-the-loop

The system operates with minimal human intervention.

This creates the greatest attribution and accountability questions.

27. Non-Human Intentionality and Evidence

In an AI dispute, the court may need evidence concerning:

source code;

system architecture;

model documentation;

prompts;

logs;

metadata;

audit trails;

access records;

electronic signatures;

system permissions;

training data;

transaction history;

human instructions;

system alerts.

Evidence chain

Human instruction → System processing → Machine output → Electronic record → Damage

The stronger the evidence chain, the easier it becomes to determine legal attribution.

28. Non-Human Intentionality and Privacy

AI systems often process personal information.

This introduces UAE data-protection considerations.

Questions include:

Was the processing lawful?

Was consent required?

Was there a legitimate legal basis?

Was excessive data collected?

Was the system secure?

Was data shared with third parties?

Was automated processing involved?

Was personal information retained appropriately?

Again, the AI does not normally become the legal controller merely because it processes data.

The relevant legal responsibility usually belongs to the organization/person identified under applicable data-protection rules.

29. Non-Human Intentionality and Algorithmic Bias

Suppose an automated system produces systematically inaccurate outcomes.

The legal analysis should not simply say:

"The algorithm was biased."

Instead:

What data were used?

Was the data accurate?

What duty existed?

What methodology was employed?

Was the outcome foreseeable?

Was the system properly tested?

Was there human review?

Did the outcome cause legally recognized damage?

Who controlled the system?

What legal rule establishes responsibility?

This converts a technological problem into a legally analyzable problem.

30. Non-Human Intentionality and Autonomous Vehicles

Consider an autonomous vehicle.

The vehicle makes a decision without immediate human intervention.

If an accident occurs, potential actors may include:

vehicle manufacturer;

software developer;

sensor manufacturer;

owner;

operator;

maintenance provider;

mapping/data provider.

The vehicle itself would not ordinarily be treated as the final civil defendant merely because it selected the driving action.

The court would investigate:

Defect → Duty → Causation → Damage → Attribution

31. Non-Human Intentionality and Legal Interpretation

A useful principle is:

Courts should distinguish technological autonomy from legal autonomy.

Technological autonomy means:

The system can operate without continuous human instructions.

Legal autonomy means:

The law recognizes the system as an independent bearer of rights and obligations.

The first does not automatically create the second.

32. Important Distinctions

Key point

Autonomy ≠ Agency ≠ Legal Personality ≠ Liability

33. Practical UAE Example

Suppose a UAE company operates an AI purchasing platform.

The AI:

searches suppliers;

negotiates automatically;

selects Supplier A;

sends an electronic purchase order;

authorizes payment;

receives defective goods.

Supplier A argues:

"The AI entered the contract, not the company."

The court would need to examine:

Step 1 — Identity

Who owns the purchasing account?

Step 2 — Authority

Who authorized the AI?

Step 3 — Contract

Were the contractual requirements satisfied?

Step 4 — Electronic evidence

What do the logs, emails and audit trails show?

Step 5 — Attribution

Can the AI's communication be attributed to the company?

Step 6 — Performance

Were the goods defective?

Step 7 — Remedies

What contractual or statutory remedies are available?

The central issue is therefore legal attribution, not whether the AI experienced a subjective intention to purchase.

34. Proposed UAE Analytical Framework

A useful framework for courts and lawyers is:

N-H-I-L Framework

N — Non-human conduct

Identify exactly what the system did.

H — Human/legal-person connection

Identify owner, controller, developer, employer, principal or authorized user.

I — Intention/authority

Determine the legally relevant human or organizational intention and authority.

L — Liability

Determine duty, breach, causation, damage and remedy.

Formula

Non-Human Conduct → Attribution → Legal Intention/Authority → Legal Rule → Liability

35. Seven Questions for Judicial Analysis

When a dispute involves AI or another non-human system, ask:

What exactly did the system do?

Who designed or deployed it?

Who authorized its operation?

Can the electronic conduct be attributed to a legal person?

What evidence proves attribution?

What duty or contractual obligation applies?

What damage and causation have been established?

These questions prevent the court from confusing technical autonomy with legal responsibility.

36. Major Challenges for UAE Civil Law

The increasing use of autonomous systems may create future questions concerning:

1. AI agency

Can an AI system legally act as an agent?

2. AI contracting

When does automated negotiation produce binding consent?

3. AI mistake

Who bears responsibility for an automated error?

4. AI negligence

Who is responsible when no person intended the harm?

5. AI evidence

How should courts verify machine-generated evidence?

6. AI causation

How should responsibility be divided among multiple technological actors?

7. AI personality

Should highly autonomous systems ever receive limited legal personality?

The last question is primarily a future policy and legislative question, not a conclusion that current UAE law already grants AI independent civil personality.

37. Exam-Oriented Case Table

Important: These are primarily DIFC/ADGM authorities and illustrative authorities for digital legal analysis. They should not be treated as binding mainland UAE precedents. Mainland UAE courts operate within the federal codified legal framework.

38. Revision Points

Remember these points:

AI is not automatically a legal person.

Automation does not automatically create legal agency.

Machine behaviour is different from human intention.

Legal analysis focuses heavily on attribution.

Electronic transactions can have legal effect without a human manually performing every step.

Corporate responsibility may arise from deployment, authorization, supervision or contractual obligations.

AI-caused harm still requires analysis of duty, breach, causation and damage.

Digital evidence can establish who controlled or authorized a system.

Algorithmic autonomy does not automatically eliminate human/legal-person responsibility.

Future recognition of AI legal personality would require a clear legal basis rather than merely technological capability.

39. Short Exam Answer

Non-human intentionality in UAE civil-law analysis concerns the legal treatment of conduct produced by AI, algorithms, automated systems, smart contracts and other non-human technologies. UAE civil law does not presently treat technological autonomy itself as creating independent legal personality. Consequently, the principal questions are attribution, authority, consent, causation, duty and damage. Electronic-transactions and evidence frameworks allow courts to recognize electronic records and automated communications, but the legal consequences normally remain connected to the natural or juridical person operating, authorizing, controlling or benefiting from the system. Cases such as ICICI Bank v Shetty, GFH Capital v Haigh, Ondina v Olin, Naho v Neukirchi, Tarig v Expresso Telecom and Jonathan Lau v Qashio illustrate the importance of electronic attribution, authority, consent and digital evidence. Therefore, the fundamental principle is:

Technological autonomy does not automatically equal legal autonomy.

40. Conclusion

Non-human intentionality is an emerging issue at the intersection of UAE civil law, electronic transactions, AI, evidence, agency, contract and civil liability.

The most important legal distinction is:

A system may act autonomously without becoming a legal person.

Accordingly, UAE civil-law analysis should move through:

Non-Human Conduct → Authentication → Attribution → Authority → Consent → Duty → Causation → Damage → Liability → Remedy

The future challenge will be determining how UAE legislation and courts should deal with increasingly autonomous AI systems. Until legislation expressly provides otherwise, the stronger analytical approach is to locate legal responsibility in the natural person or juridical person connected with the system, rather than assuming that the machine itself possesses independent legal intention or civil personality.