Civil Law and UAE: Probabilistic Reasoning in Liability Assessment

1. Meaning

Probabilistic reasoning in liability assessment means the court evaluates the available evidence and determines whether a fact, event, breach, causal connection, or loss is more likely than not to have occurred.

It becomes important when the evidence does not provide absolute certainty.

Typical situations include:

several possible causes of an accident;

medical or scientific uncertainty;

professional negligence;

financial loss caused by multiple factors;

technical or construction failures;

cyber incidents;

insurance claims;

loss of an opportunity;

multiple persons contributing to one injury;

disputes involving indirect or circumstantial evidence.

The important point is that probabilistic reasoning is an evidentiary method, not a separate UAE category of civil liability. The court still has to identify the applicable legal rule and establish the required elements of liability.

Under the current Federal Decree-Law No. 25 of 2025 promulgating the Civil Transactions Law, effective from 1 June 2026, the harmful-act provisions establish the current federal framework. Article 246 provides the basic compensation rule for harmful acts; Article 247 distinguishes direct harm from harm occurring by causation; Article 249 addresses external causes such as force majeure, third-party acts and the injured person's own act. (LEXAI)

2. Basic Legal Structure

The UAE civil-liability analysis can be simplified as:

Conduct
↓
Legally relevant responsibility/fault
↓
Damage
↓
Causal connection
↓
Evidence
↓
Probability assessment
↓
Liability
↓
Compensation

Therefore, probabilistic reasoning generally operates inside the process of proving liability.

It does not mean:

“There is a 40% possibility of liability, so the defendant pays 40%.”

Rather, the court may ask:

“Having considered all the evidence, is it more probable than not that the defendant's conduct caused the relevant loss?”

3. Current UAE Statutory Foundation

Article 245

The harmful-act chapter applies to civil liability arising from harmful acts by natural and legal persons, subject to special legislation.

Article 246

The general principle is that a harmful act causing damage gives rise to an obligation to compensate.

Article 247

The law distinguishes between:

direct harm, and

harm caused through causation.

For causative harm, the statutory provision considers circumstances such as wrongful conduct, intentional conduct, or conduct leading to the harm. (LEXAI)

Article 249

An external cause beyond the person's control may exclude compensation, including:

force majeure;

sudden accident;

act of a third party;

act of the injured person,

subject to the statutory qualification that law or agreement may provide otherwise. (LEXAI)

This is important for probabilistic reasoning because the court may need to compare the defendant's explanation with alternative causes.

4. Probability Is Not Speculation

There is an important difference between probabilistic reasoning and speculation.

Probabilistic reasoning

The evidence makes one explanation more probable than competing explanations.

Speculation

The defendant's conduct could possibly have caused the loss.

The second proposition is normally insufficient.

The court therefore considers:

Evidence → inference → probability → legal conclusion

rather than:

Possibility → assumption → liability.

5. Balance of Probabilities

DIFC jurisprudence, which is particularly useful as a comparative UAE authority, expressly recognises the balance of probabilities as the civil standard.

In Graciela Ltd v Giacobbe, the Court explained that an event is established where, on the evidence, it is more likely than not to have occurred. The Court also relied on circumstantial evidence in determining responsibility for an IT-system attack. (DIFC Courts)

Thus, probabilistic reasoning does not require mathematical certainty.

For example:

Explanation A: 70% supported by evidence

Explanation B: 30% supported by evidence

The court does not necessarily award “70% liability.”

Instead, if the applicable civil standard is satisfied, it may find A established as a matter of fact.

6. Case Law 1 — Graciela Ltd v Giacobbe [2014] DIFC CFI 027

This is an important authority for probabilistic reasoning.

The claimant alleged that a former employee had sabotaged its IT system. There was substantial circumstantial evidence rather than a simple direct eyewitness account.

The Court considered:

access to the system;

knowledge of passwords and IP information;

creation of a secret server;

copying of data;

deletion of information;

the defendant's conduct after the attack.

The Court concluded that the civil standard was satisfied on the balance of probabilities. (DIFC Courts)

Principle

A liability finding may be based on the combined weight of circumstantial evidence.

This is particularly important for:

cyber liability;

fraud;

concealment;

technical misconduct;

corporate wrongdoing.

7. Case Law 2 — AES Middle East Insurance Broker LLC v GSB Capital Ltd [2023] DIFC CFI 060

This case provides an especially useful explanation of collective evidential reasoning.

The claimants relied upon circumstantial evidence and asked the Court to draw inferences concerning alleged wrongdoing.

The Court stated that individual facts do not necessarily have to be proved independently in isolation if the combined weight of the evidence establishes the relevant inference on the balance of probabilities. (DIFC Courts)

Principle

The court may examine:

Fact A + Fact B + Fact C + Fact D

and conclude:

Overall inference = more probable than the competing explanation.

This is a central form of probabilistic reasoning.

8. Case Law 3 — Amira C Foods International DMCC v IDBI Bank Ltd [2018] DIFC CFI 027

This is one of the most useful authorities on probabilistic causation.

The Court considered whether negligence had caused the claimed loss and emphasised that the court should examine alternative theories of causation.

Where the defendant's alternative explanations are, on balance, improbable, that may strengthen the inference that the claimant's explanation is correct. (DIFC Courts)

Principle

The reasoning can be represented as:

Claimant's explanation + proven negligence + resulting damage

versus

Defendant's alternative explanations

If the claimant's explanation is more probable after the competing explanations have been tested, causation may be established.

Important limitation

The case does not create an automatic presumption of causation.

The burden of proof is not simply transferred to the defendant.

9. Case Law 4 — Haya Spa LLC v Harper Real Estate / Hasan Real Estate [2016] DIFC SCT 150

The Court considered the causal requirements under the DIFC Law of Obligations.

The relevant framework required the claimant to establish that:

but for the defendant's conduct, the loss would not have occurred; and

the conduct was a substantial cause of the loss.

The Court also considered whether a later event had become a supervening event, thereby breaking the operative causal connection. (DIFC Courts)

Principle

Probabilistic reasoning must consider the entire causal chain.

For example:

A's negligence → initial loss → later event → additional loss

The defendant may be responsible for the initial loss but not necessarily for all subsequent losses.

10. Case Law 5 — Ludiala v Lucaan Ltd [2020] DIFC SCT 139

This case demonstrates the opposite situation: insufficient evidence of causation.

The claimant alleged psychological injury resulting from workplace conduct.

The Court examined the medical material and concluded that it did not sufficiently establish that the defendant's conduct caused the psychiatric harm. The claimant's own assertion of causation was insufficient, and the fact that the condition appeared after employment began did not itself establish causation. (DIFC Courts)

Principle

Sequence is not causation.

The fact that:

Event A occurred before Injury B

does not automatically prove:

Event A caused Injury B.

This is particularly important in:

medical negligence;

psychiatric injury;

occupational injury;

environmental claims.

11. Case Law 6 — Oheo Bank v Parker [2025] DIFC CA 006

This is a recent authority concerning a finding that a regulatory breach was a material cause of loss.

The underlying tribunal found, on the balance of probabilities, that the bank's failure to explain the purpose of an indemnity was a material cause of the claimant's loss. The Court of Appeal considered the resulting challenge to the award. (DIFC Courts)

Principle

A defendant's conduct does not necessarily need to be:

the only cause

of the loss.

It can be a material cause within a larger factual chain.

Thus:

Cause A + Cause B + Cause C → Loss

may still permit legal responsibility for Cause A where the applicable legal test is satisfied.

12. Case Law 7 — Alawwal Capital JSC v Rasmala Investment Bank Ltd [2023] DIFC CFI 038

This case concerned alleged negligent representations relating to an investment.

The claimant argued that, had the representations not been made, it would not have invested in the relevant fund and would instead have chosen safer investments.

The Court applied the DIFC Law of Obligations' causation framework, including the but-for test and substantial causation. (DIFC Courts)

Importance

This demonstrates probabilistic reasoning in a counterfactual situation.

The court asks:

What would probably have happened if the defendant had acted correctly?

This is common in:

negligent financial advice;

investment claims;

professional negligence;

lost opportunities.

13. Case Law 8 — Qatar General Insurance & Reinsurance Co QSPC v Emrgent Risk Solutions Ltd [2024] DIFC CFI 053

This case illustrates probabilistic reasoning concerning a hypothetical outcome.

The Court considered what would probably have happened if the relevant insurance intermediary had acted properly, including the likelihood of an insurance recovery.

The analysis was conducted using the civil standard of proof—the balance of probabilities. (DIFC Courts)

Principle

Courts may sometimes have to assess:

What probably would have happened?

rather than merely:

What actually happened?

This is especially important in damages assessment.

14. Case Law 9 — BAM Higgs & Hill LLC v Affan Innovative Structures LLC [2021] DIFC CFI 106

The Court referred to Dubai Commercial Appeal 445/2020/1034 and emphasised that liability requires the coexistence of:

fault;

damage; and

causal connection.

The cited Dubai authority states that if one of those elements is absent, the liability claim fails. (DIFC Courts)

Principle

Probabilistic reasoning cannot eliminate the requirement for a legally recognised causal connection.

A court cannot reason:

“The defendant was negligent, therefore the defendant must pay.”

It must still determine:

What damage did that conduct legally cause?

15. Case Law 10 — LALS Holdings Ltd v Emirates Insurance Co & SIACI Insurance Brokers [2022] DIFC CFI 073

The Court dealt with insurance coverage questions and expressly recognised that the burden of proof rested on the claimants, with the applicable standard being the balance of probabilities. (DIFC Courts)

Principle

The probability assessment depends on the evidence available in the particular case.

There is no universal formula such as:

“Expert evidence automatically establishes liability.”

Instead, the court considers the entire evidentiary record.

16. Probabilistic Reasoning in Multiple-Cause Cases

Suppose a building collapses.

Possible causes include:

defective design;

defective materials;

negligent construction;

unusual weather;

poor maintenance;

subsequent alterations.

The court may have to determine:

Which cause or causes probably produced the collapse?

The analysis could be:

The court does not simply count the number of causes.

It weighs their evidential strength and causal relevance.

17. Probabilistic Reasoning and Alternative Causes

This is one of the most important aspects.

A claimant says:

“The defendant caused the damage.”

The defendant responds:

“The damage resulted from another cause.”

The court must examine both propositions.

The Amira C Foods approach is particularly useful: alternative theories should be considered before determining where the probability lies. (DIFC Courts)

Therefore:

Claimant's theory

Defendant's alternative theory

Expert evidence

Documents

Circumstantial evidence

Chronology

= Overall probability assessment

18. Probabilistic Reasoning and Contributory Conduct

The claimant may also have contributed to the damage.

For example:

Defendant's negligence → 70% causal contribution

Claimant's conduct → 30% causal contribution

But it is important not to confuse:

Causation

Did the defendant's conduct contribute to the harm?

with:

Allocation

What legal consequence follows from the claimant's contribution?

The current UAE Civil Transactions Law separately addresses external causes and the consequences of conduct by the injured party. Article 249 is therefore particularly relevant where the claimant's own conduct is alleged to have caused or contributed to the harm. (LEXAI)

19. Probabilistic Reasoning and Intervening Events

A later event can alter the liability analysis.

Example:

Negligent act

↓

Initial injury

↓

Independent third-party act

↓

Additional injury

The court may conclude that the original defendant caused the initial injury but that the later independent event caused the additional damage.

The current UAE law expressly recognises external causes, including the act of a third party and the act of the injured person, as potentially excluding compensation. (LEXAI)

20. Probabilistic Reasoning in Medical Liability

Medical disputes are particularly difficult.

Suppose:

Patient has underlying disease

Doctor allegedly makes a negligent decision

Patient subsequently deteriorates

There may be several explanations:

natural progression;

pre-existing condition;

negligent treatment;

medication;

delay;

unrelated complication.

The court therefore needs reliable medical evidence.

Ludiala v Lucaan demonstrates that medical records must actually support the causal connection; merely recording the claimant's belief that workplace conduct caused the condition is insufficient. (DIFC Courts)

21. Probabilistic Reasoning in Professional Liability

Professionals may face claims involving:

lawyers;

accountants;

engineers;

architects;

financial advisers;

insurance brokers;

consultants.

The claimant may need to establish:

Professional error → probable consequence → actual loss

For example:

If an accountant had correctly identified a tax problem, would the claimant probably have avoided the loss?

This is a counterfactual question.

The court may therefore examine:

what the professional should have done;

what the claimant would probably have done;

whether the alternative action would probably have prevented the loss.

22. Probabilistic Reasoning in Financial Liability

Financial cases often involve uncertainty.

Suppose a bank negligently provides incorrect information.

The claimant loses AED 10 million.

But several market events also occurred.

The court may need to determine:

Was the loss caused by the bank's conduct or by market conditions?

Evidence may include:

investment records;

market data;

expert reports;

previous investment behaviour;

contemporaneous communications;

alternative investment opportunities.

Alawwal Capital v Rasmala illustrates this type of counterfactual causation analysis. (DIFC Courts)

23. Probabilistic Reasoning in Insurance

Insurance disputes frequently involve questions such as:

Was the insured event the probable cause of the loss?

or:

Would the loss probably have occurred even without the alleged breach?

LALS Holdings demonstrates that the court applies the balance of probabilities to insurance issues while placing the evidentiary burden on the claimant. (DIFC Courts)

Thus, probabilistic reasoning is particularly important where:

several risks overlap;

policy exclusions are disputed;

business interruption is claimed;

alternative causes exist.

24. Probabilistic Reasoning in Cyber Liability

Digital disputes often lack a direct witness.

For example:

Cyberattack → data corruption → financial loss

Possible causes:

employee misconduct;

external hacker;

software vulnerability;

vendor failure;

inadequate cybersecurity;

user error.

Graciela v Giacobbe demonstrates the usefulness of circumstantial evidence in an IT-system dispute. (DIFC Courts)

Relevant evidence may include:

server logs;

access records;

IP information;

metadata;

system backups;

timestamps;

security reports.

The court can construct a probability-based factual chain from this material.

25. Probabilistic Reasoning and AI Liability

The concept is increasingly relevant to AI-related disputes.

Suppose:

AI system → incorrect recommendation → financial loss

Potential causes could include:

defective algorithm;

inadequate training data;

incorrect deployment;

human input;

inadequate supervision;

third-party data;

cyber manipulation.

The court may need to determine:

Which factor probably caused the particular loss?

The AI system itself does not eliminate the ordinary requirements of legal attribution, damage and causation.

The current UAE harmful-act framework's distinction between direct and causative harm is therefore potentially relevant to technologically complex liability disputes. (LEXAI)

26. Role of Expert Evidence

Probabilistic liability assessment frequently depends on expert evidence.

Medical expert

Determines likely medical causation.

Engineering expert

Determines likely technical cause.

Financial expert

Models alternative financial outcomes.

IT expert

Analyses system failure or cyberattack.

Insurance expert

Assesses coverage and causal loss.

However:

The expert provides evidence; the court decides legal responsibility.

The court may accept, reject or give limited weight to expert evidence after considering the entire record.

27. Circumstantial Evidence

Probabilistic reasoning is particularly important when direct evidence is unavailable.

The court can consider:

conduct before the event;

conduct after the event;

access;

opportunity;

timing;

technical evidence;

financial records;

inconsistencies;

documentary evidence;

expert evidence.

Graciela is a strong illustration: the Court reached its conclusion using a combination of circumstances concerning access, hidden infrastructure, data copying and subsequent conduct. (DIFC Courts)

28. Probability and Serious Allegations

The civil standard remains the balance of probabilities, even where allegations are serious.

In Graciela, the Court explained that seriousness affects the evaluation of inherent probability and the strength of evidence needed to reach satisfaction, but does not create a separate higher civil standard. (DIFC Courts)

Thus:

Serious allegation ≠ criminal standard automatically

but:

Serious allegation → court carefully evaluates the evidential probabilities.

29. Probability and Damages

Probabilistic reasoning is not restricted to establishing liability.

It can also affect quantification of loss.

For example:

A negligent professional caused a 40% probability of losing a commercial opportunity worth AED 5 million.

The legal analysis may distinguish:

whether the opportunity existed;

whether the defendant caused its loss;

the probability that the opportunity would have materialised;

the appropriate valuation.

This is different from saying that the defendant is automatically “40% liable.”

30. Liability Assessment vs Mathematical Probability

A crucial examination point:

A court may use statistical or expert evidence, but the legal conclusion remains a judicial determination.

31. Important Limits

Probabilistic reasoning should not be used to:

1. Reverse the burden of proof automatically

The claimant generally remains responsible for establishing the required elements.

2. Treat possibility as proof

A possible cause is not necessarily a proven cause.

3. Ignore alternative explanations

Competing causes must be considered where relevant.

4. Replace legal causation with statistics

Statistical association alone does not necessarily establish legal causation.

5. Treat expert opinion as conclusive

Experts assist the court; they do not decide liability.

6. Assume negligence equals damage

Fault and causation remain distinct questions.

7. Ignore intervening causes

A later independent event may break the causal chain.

32. Ten Important Cases — Quick Table

33. Federal UAE vs DIFC Authorities

A distinction is essential.

UAE Mainland

The principal current framework is:

Federal Decree-Law No. 25 of 2025 — Civil Transactions Law

effective from 1 June 2026. Articles 245–255 are particularly relevant to harmful acts, causation and compensation. (LEXAI)

DIFC

DIFC has its own legal framework, including the DIFC Law of Obligations.

Therefore, cases such as Haya Spa, Amira C Foods, Graciela and Oheo Bank are valuable UAE/DIFC comparative authorities, but they should not be described as automatically binding interpretations of the federal Civil Transactions Law.

34. Practical Liability Formula

For an examination or legal analysis, use:

PROBABILISTIC LIABILITY ASSESSMENT

1. Identify the conduct

↓

2. Identify the applicable legal duty/rule

↓

3. Establish actual damage

↓

4. Identify the alleged causal connection

↓

5. Examine direct and circumstantial evidence

↓

6. Consider alternative causes

↓

7. Consider intervening events

↓

8. Assess the evidence on the balance of probabilities

↓

9. Determine legal responsibility

↓

10. Quantify compensable loss

35. Example

Assume a contractor constructs a building.

Six months later, serious cracks appear.

Three possible causes exist:

defective materials;

negligent construction;

unusually heavy rainfall.

The claimant produces an engineering report stating that defective construction is the most probable cause.

The contractor argues that rainfall caused the cracks.

The court would examine:

construction records;

material quality;

weather records;

photographs;

expert reports;

timing;

maintenance;

alternative explanations.

The court does not need absolute scientific certainty.

It must determine whether the evidence establishes the legally relevant causal connection according to the applicable standard.

This is probabilistic reasoning in liability assessment.

36. Conclusion

Probabilistic reasoning in UAE liability assessment is the process through which courts evaluate uncertain, competing or circumstantial evidence to determine whether the elements required for civil responsibility have been established.

The current UAE Civil Transactions Law provides the federal foundation by recognising harmful acts, distinguishing direct and causative harm, addressing external causes, and linking compensation to legally relevant loss. (LEXAI)

The DIFC cases provide particularly clear illustrations of the methodology. Graciela demonstrates circumstantial proof; AES v GSB demonstrates the combined weight of evidence; Amira C Foods demonstrates evaluation of competing causal theories; Haya Spa demonstrates operative causation and intervening events; Ludiala shows that insufficient medical evidence cannot establish causation; and Oheo Bank demonstrates that a breach can be a material cause of loss without necessarily being the only cause. (DIFC Courts)

One-line revision principle

Probabilistic reasoning in UAE civil liability does not mean liability based on mere possibility; it means reaching a legally sufficient conclusion from the totality of evidence, competing explanations and causal circumstances according to the applicable civil standard of proof.