Civil Law and UAE: Probabilistic Causation in Tort Law

1. Meaning of Probabilistic Causation

Probabilistic causation arises where the claimant cannot establish the causal mechanism with absolute scientific certainty but argues that, on the available evidence, the defendant's act was more likely than not to have caused the damage.

It is especially relevant where:

• several possible causes exist;
• scientific evidence is incomplete;
• medical causation is uncertain;
• multiple persons contributed to the harm;
• technical failures have several possible explanations;
• the precise mechanism of damage cannot be reconstructed;
• future or opportunity losses depend on probabilities.

A crucial point under UAE law is that “probabilistic causation” is not a separate statutory tort doctrine with a fixed percentage test. Rather, it is better understood as an evidential problem within the ordinary requirement of proving the causal connection between wrongful conduct and damage.

The current Federal Decree-Law No. 25 of 2025 Civil Transactions Law came into force on 1 June 2026. Article 247 now expressly distinguishes direct harm from harm occurring by causation and provides different consequences depending upon the causal circumstances. Article 249 further recognises external causes such as force majeure, an act of a third party or an act of the injured person as potentially breaking liability.

2. Basic UAE Tort Formula

The basic structure can be represented as:

Wrongful Act / Fault
↓
Damage
↓
Causal Connection
↓
Attribution of Damage to Defendant
↓
Compensation

The causal connection is essential.

A defendant may have acted negligently, but:

Negligence without causally connected damage does not automatically produce tort liability.

Dubai Commercial Appeal 445/2020/1034, as reproduced by the DIFC Courts in BAM Higgs & Hill v Affan, states that liability requires fault, damage and a causal connection, and that absence of any one of those elements defeats liability.

3. Current UAE Civil Transactions Law

The current statutory framework is particularly important because the 2025 Civil Transactions Law is now in force.

Article 246

Every act causing harm creates an obligation to compensate.

Article 247

The new law expressly distinguishes:

1. direct harm, and
2. harm occurring by causation.

Where harm occurs through causation, liability depends upon the statutory requirements concerning transgression, intention or the causal act. Where a direct actor and a causal actor combine, the statutory rule generally attributes the consequence to the direct actor.

Article 249

A person may escape compensation where the harm resulted from an external cause beyond their control, including:

• act of God;
• sudden accident;
• force majeure;
• act of a third party;
• act of the injured person,

subject to the qualification that law or agreement may provide otherwise.

This is highly relevant to probabilistic causation because alternative causes may weaken or defeat the claimant's causal case.

4. “More Likely Than Not” Rather Than Absolute Certainty

Civil litigation normally does not require mathematical certainty.

The central question is:

On the evidence as a whole, is the defendant's conduct more probably than not a cause of the claimant's loss?

This is particularly clear in DIFC jurisprudence. In Graciela Ltd v Giacobbe, the DIFC Court stated that the civil standard is the balance of probabilities—whether an event was more likely than not to have occurred.

Therefore, evidence can be:

• circumstantial;
• inferential;
• expert-based;
• statistical;
• documentary;
• chronological.

The absence of direct evidence does not necessarily defeat causation.

5. Case Law

Case 1 — BAM Higgs & Hill LLC v Affan Innovative Structures LLC [2021] DIFC CFI 106

This is an important UAE/DIFC authority on causation.

The Court referred to Dubai Commercial Appeal 445/2020/1034 and explained that contractual or tortious liability requires:

1. fault;
2. damage; and
3. causal connection.

The case also emphasises that the causal link is an essential component of liability.

Principle

Probabilistic evidence cannot substitute for causation altogether.

The claimant still has to establish a legally sufficient connection between the defendant's conduct and the damage.

6. Case 2 — Haya Spa LLC v Harper Real Estate & Hasan Real Estate [2016] DIFC SCT 150

This is one of the clearest authorities concerning the actual causal test.

The Court applied Articles 10 and 11 of the DIFC Law of Obligations and required the claimant to establish that:

• but for the defendant's conduct, the loss would not have occurred; and
• the defendant's conduct was a substantial cause of the loss.

The Court also considered whether subsequent events constituted a supervening event that broke the causal connection.

Importance

This provides a useful model for analysing probabilistic causation:

Defendant's conduct → but-for connection → substantial cause → no effective intervening cause.

The case also demonstrates that causation may change over time. A defendant's conduct may be causal during one period but cease to be an operative cause after a later event.

7. Case 3 — Ludiala v Lucaan Limited [2020] DIFC SCT 139

The claimant alleged psychological harm arising from workplace conduct.

The Court held that causation had not been sufficiently established. Medical evidence did not adequately establish that the defendant's conduct caused the claimant's psychiatric condition. The fact that the claimant had not suffered mental illness before the employment was insufficient by itself.

Principle

Temporal sequence is not the same as legal causation.

The reasoning can be expressed as:

“It happened after the defendant's conduct” ≠ “it was caused by the defendant's conduct.”

This is especially important in medical negligence and psychiatric injury claims.

8. Case 4 — Amira C Foods International DMCC v IDBI Bank Ltd [2018] DIFC CFI 027

This case is particularly relevant to inference-based causation.

The Court considered an argument based on the principle that where negligence has been established and the resulting loss is of the type that would ordinarily be expected from that negligence, causation may sometimes be inferred even though the claimant cannot demonstrate the precise physical or commercial mechanism.

The Court discussed the importance of examining alternative explanations and determining where the probability lies.

Importance

This is very close to the concept of probabilistic causation.

The claimant does not necessarily need to prove:

“This is the precise mechanical chain through which the damage occurred.”

It may be sufficient, depending on the applicable law and evidence, to establish:

“The defendant was negligent, this type of damage followed, and competing explanations are insufficient to displace the inference that the negligence caused it.”

However, this is not a reversal of the burden of proof.

9. Case 5 — Oheo Bank v Parker [2025] DIFC CA 006

The Court of Appeal considered an arbitral finding that a bank's regulatory breach was a material cause of the claimant's loss.

The underlying tribunal had concluded, on the balance of probabilities, that the bank's failure to explain the purpose of an indemnity was a material cause of the loss.

Principle

A defendant's conduct does not necessarily have to be the only cause.

A causal contribution can be sufficient where the applicable legal test is satisfied.

Thus:

Cause A + Cause B + Cause C → Loss

does not necessarily mean that none of the causes is legally relevant.

10. Case 6 — Qatar General Insurance & Reinsurance Company QSPC v Emrgent Risk Solutions Ltd [2024] DIFC CFI 053

This case provides an especially useful example of probabilistic assessment of hypothetical events.

The Court considered what would probably have happened if the insurance broker had not breached its duty. The assessment involved determining the likelihood that an insurer would have made a payment.

The Court described the assessment as one conducted on the balance of probabilities, including assessment of the likelihood of the hypothetical insurance recovery.

Importance

This demonstrates a second form of probabilistic reasoning:

Historical causation

Did the defendant's conduct cause the actual loss?

and

Counterfactual causation

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

Both can require probability-based reasoning.

11. Case 7 — Graciela Limited v Giacobbe [2014] DIFC CFI 027

The defendant was accused of sabotaging an IT system.

There was no simple direct eyewitness account of the sabotage. The Court relied heavily upon circumstantial evidence, including the defendant's knowledge, access and conduct.

The Court applied the ordinary civil standard of proof—the balance of probabilities.

Principle

Causation and responsibility may sometimes be established through a combination of circumstances.

This is particularly relevant to:

• cyber torts;
• digital evidence;
• algorithmic failures;
• technical accidents;
• confidential information cases.

The claimant does not necessarily need a single piece of evidence labelled “causation.”

12. Case 8 — Alawwal Capital JSC v Rasmala Investment Bank Ltd [2023] DIFC CFI 038

The claimant alleged that negligent representations caused it to invest in a fund when it otherwise would have selected safer investments.

The Court identified Article 10 of the DIFC Law of Obligations as establishing the but-for and substantial-cause test.

Principle

Probabilistic causation can arise where the court must determine:

What would the claimant probably have done if the defendant had acted correctly?

This is particularly relevant to:

• financial mis-selling;
• negligent advice;
• investment decisions;
• professional negligence;
• lost opportunities.

13. Case 9 — Shihab Khalil v Shuaa Capital PSC [2009] DIFC CFI 017

The Court explained that a negligence claim requires both:

1. want of due care; and
2. proof that the carelessness caused loss.

If the claimant fails to establish the causal component, the tort claim cannot succeed.

Principle

Causation is an independent element of the tort.

It cannot simply be presumed from the existence of negligent conduct.

14. Case 10 — Faizal Babu Moorkath v Expresso Telecom Group Ltd [2023] DIFC CFI 008

The Court stressed that the claimant must first establish actionable loss before causation can properly be considered. It rejected the claimed loss on the evidence and then addressed duty, breach and causation as an alternative analysis.

Principle

The causal question is:

What legally recognised loss was actually caused?

A claimant cannot establish causation merely by identifying some undesirable consequence.

15. Probabilistic Causation and Medical Negligence

Medical cases frequently create difficult causal questions.

For example:

A patient suffers a medical condition after treatment.

There may be several possible causes:

• underlying disease;
• natural progression;
• treatment;
• delay;
• negligent diagnosis;
• medication;
• another medical condition.

The claimant may argue:

“The negligent treatment materially increased the probability of the injury.”

The court must distinguish:

A. Increased risk

The defendant's conduct increased the risk of harm.

B. Actual causation

The defendant's conduct probably caused the particular harm suffered.

These are not necessarily identical.

Ludiala v Lucaan demonstrates the evidentiary difficulty: medical records that merely describe symptoms or repeat the claimant's own attribution may not establish the required causal connection.

16. Multiple Causes

Suppose damage results from:

• negligent construction;
• defective materials;
• extraordinary weather;
• claimant's own conduct.

The court may have to determine:

Which causes actually contributed to the damage?

The current Civil Transactions Law expressly recognises external causes and also addresses situations where direct and causal actors combine.

Therefore, probabilistic reasoning can help identify:

• primary cause;
• contributing cause;
• intervening cause;
• external cause;
• claimant's contribution.

17. Intervening Cause

An intervening event may break the causal chain.

Example:

Defendant's negligence → initial injury → independent extraordinary event → additional damage

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

Haya Spa v Harper Real Estate illustrates this approach: the Court examined whether subsequent events had severed the operative causal connection.

The current UAE law's Article 249 similarly recognises external causes such as force majeure, third-party acts and acts of the injured person as potentially excluding compensation.

18. Contributory Conduct of the Victim

Probabilistic causation can become complicated when the claimant's own conduct contributed to the loss.

For example:

Defendant's negligence = 60% contribution

Claimant's conduct = 40% contribution

The legal consequence depends upon the applicable statutory framework and the circumstances.

The important conceptual point is that causation and apportionment are different questions:

Causation

Did the defendant contribute to the damage?

Apportionment

What portion of the legally recoverable damage should be attributed to that defendant?

19. Loss of Chance and Probability

A particularly sophisticated application occurs when the claimant has lost an opportunity.

Example:

A financial adviser negligently provides information.

The client says:

“If I had received correct advice, I probably would have invested differently.”

The court must assess a counterfactual.

The question becomes:

What was the probability that the claimant would have taken the alternative course?

The DIFC framework expressly recognises compensation for loss of an opportunity in proportion to the probability of its occurrence. The principle was discussed in Amira C Foods.

This is an important distinction:

Probability can be used to quantify an opportunity loss without necessarily converting probability itself into proof that a physical injury occurred.

20. Difference Between Probabilistic Causation and Probability of Loss

These concepts should not be confused.

21. Role of Experts

In difficult causation cases, expert evidence may be critical.

Examples include:

Medical expert

Whether treatment probably caused injury.

Engineering expert

Whether a construction defect probably caused structural failure.

Financial expert

Whether a particular representation probably caused an investment loss.

IT expert

Whether a cybersecurity failure caused the data loss.

Environmental expert

Whether pollution probably caused the physical damage.

But:

An expert does not decide legal causation.

The expert provides technical evidence. The court determines the legal significance of that evidence.

22. Statistical Evidence

Statistical evidence may be useful where individual causation is difficult.

For example:

• failure rates;
• accident frequencies;
• medical risk ratios;
• system error rates;
• historical investment patterns;
• engineering failure probabilities.

However:

Statistical association ≠ automatic legal causation.

A statistical correlation may support a causal inference, but the court still has to examine the particular claimant's circumstances.

23. AI and Algorithmic Tort Claims

Probabilistic causation is becoming particularly important in AI-related civil liability.

Suppose:

AI system → incorrect decision → financial loss

There may be several potential causes:

• defective training data;
• algorithmic design;
• human supervision;
• incorrect deployment;
• user input;
• third-party data;
• cyberattack.

The claimant must identify the causal connection.

The same framework can be represented as:

AI Error → Actual Harm → Evidence → Alternative Causes → Probability → Legal Causation

The current UAE Civil Transactions Law's distinction between direct harm and harm occurring by causation is potentially relevant to such technologically complex disputes.

24. Digital Evidence and Probabilistic Causation

Digital disputes often involve indirect evidence.

For example:

• server logs;
• access records;
• timestamps;
• metadata;
• system alerts;
• audit trails;
• IP information;
• blockchain records;
• transaction records.

A court can potentially construct a causal chain from these circumstances.

Graciela v Giacobbe is useful by analogy because the Court relied upon circumstantial evidence to determine responsibility for an IT-system attack.

25. Important Limit: Probability Is Not Speculation

There is a major difference between:

Probabilistic reasoning

“The evidence makes explanation A more probable than explanation B.”

and:

Speculation

“Explanation A is possible, therefore the defendant caused the damage.”

The latter is insufficient.

A claimant should ideally establish:

Evidence → inference → probability → causal conclusion

rather than:

Possibility → assumption → liability.

26. Six Essential Questions for a UAE Court

In a difficult causation case, the following questions are particularly useful:

1. What wrongful act occurred?
2. What actual damage occurred?
3. What causal mechanism is alleged?
4. What evidence supports that mechanism?
5. Are there credible alternative causes?
6. On the totality of the evidence, has the required causal connection been established?

27. Case-Law Summary

These are primarily DIFC authorities, not automatically binding precedents for UAE mainland courts. They are particularly useful for explaining evidentiary and analytical approaches to causation. The federal 2025 Civil Transactions Law is now the governing mainland statutory framework.

28. Probabilistic Causation vs Traditional Causation

29. Exam-Oriented Legal Formula

PROBABILISTIC CAUSATION =

Wrongful Act

• Actual Damage
• Evidence of Causal Connection
• Balance of Probabilities
• Evaluation of Alternative Causes
• No Effective Intervening Cause
• Attribution of Loss
  = Potential Tort Liability

30. Conclusion

Probabilistic causation in UAE tort law does not mean that a claimant can obtain damages merely by demonstrating that the defendant's conduct might have caused the harm. The claimant must establish the legally required causal connection using the applicable standard of proof.

The important idea is that civil causation does not normally require scientific certainty. Courts may draw reasonable inferences from direct and circumstantial evidence and assess competing explanations according to the civil standard. DIFC cases such as Haya Spa, Amira C Foods, Ludiala, Graciela and Oheo Bank illustrate different aspects of this process.

Under the current UAE Civil Transactions Law, the statutory distinction between direct harm and harm caused through causation, together with the rules concerning external causes, provides the current federal foundation for analysing causal responsibility.

One-line revision principle

In UAE tort law, probabilistic causation means establishing the causal connection through the totality of evidence and legally permissible inference, rather than requiring absolute scientific certainty; mere possibility or temporal sequence is insufficient.