Algorithmic Reimbursement Freeze Liability in Switzerland

(Detailed Legal Explanation + 6 Key Swiss Case Law Principles)

Switzerland does not yet have a specific statutory doctrine called “algorithmic reimbursement freeze liability.” However, the issue is increasingly relevant due to:

• Automated banking compliance systems (AML/KYC software)
• AI-driven fraud detection systems freezing payments
• FinTech platforms blocking or reversing transactions
• Insurance reimbursement engines rejecting claims automatically
• Crypto exchanges using algorithmic compliance filters

In Swiss law, liability is assessed through general civil law (Swiss Code of Obligations), banking law principles, data protection law, and Federal Supreme Court jurisprudence (Bundesgericht / BGE rulings).

The core legal question is:

When an automated system wrongly freezes or denies reimbursement, who is legally responsible—the bank/company, or the algorithmic system itself?

Swiss law is clear:
Algorithms cannot be liable; humans and institutions remain fully liable.

I. Core Legal Framework in Switzerland

Algorithmic reimbursement freeze disputes are assessed under:

• Swiss Code of Obligations (CO)
  • Art. 97 CO (contractual liability)
  • Art. 41 CO (tort liability)
  • Art. 62 CO (unjust enrichment / wrongful retention)
• Swiss Banking Act & FINMA regulations
• Swiss Federal Act on Data Protection (FADP)

Key principle:

Automation does NOT reduce legal responsibility.

II. 6 Key Swiss Case Law Principles (Federal Supreme Court Jurisprudence)

Below are six major Swiss Federal Supreme Court (BGE) jurisprudential lines that directly govern liability in algorithmic or automated reimbursement freeze scenarios.

1. Bank Duty of Care in Payment Processing (BGE jurisprudence on banking negligence)

Principle:

Swiss banks owe strict contractual and professional duty of care in executing client transactions.

Legal rule established:

If a payment is incorrectly blocked or reversed—even by internal systems—the bank is liable for damages under Art. 97 CO.

Relevance to algorithmic freezes:

• Automated fraud filters do not excuse wrongful blocking
• System errors are legally treated as organizational failures

2. Unauthorized Transaction Reversal Liability (Swiss Federal Supreme Court case law on payment execution errors)

Principle:

A bank that reverses or freezes a payment without proper legal basis breaches contract.

Legal rule:

Even “good faith” algorithmic suspicion is insufficient justification unless supported by law (e.g., AML obligations).

Relevance:

• AI fraud detection cannot override contractual payment obligations
• Wrongful freeze triggers restitution + damages

3. Strict Liability for Internal System Failures (BGE jurisprudence on organizational responsibility)

Principle:

Swiss courts consistently hold that companies are responsible for internal operational systems.

Legal rule:

Technical malfunction—whether human or automated—is attributed to the institution.

Relevance:

• Algorithmic decision systems are legally “internal tools”
• No defense of “automation error” exists

4. Unjust Enrichment in Wrongful Payment Retention (Art. 62 CO case line)

Principle:

If funds are withheld or frozen without valid legal cause, retention becomes unjust enrichment.

Legal rule:

Even temporary algorithmic freezes can trigger repayment obligations if unjustified.

Relevance:

• Reimbursement delays caused by automated compliance filters may require compensation
• Interest may accrue on wrongly withheld funds

5. Data Protection & Automated Decision-Making Restrictions (FADP jurisprudence interpretation)

Principle:

Swiss data protection law limits fully automated decisions that significantly affect individuals.

Legal rule:

If an automated system makes a financial decision (like reimbursement denial), the individual must have:

• explanation rights
• human review option

Relevance:

• Pure “black box” reimbursement denial is legally vulnerable
• Lack of transparency can make freeze unlawful

6. Duty to Verify Suspicious Transaction Freezes (Anti-Money Laundering jurisprudence)

Principle:

Under AML obligations, banks must freeze suspicious transactions—but also act proportionately.

Legal rule:

Courts require:

• reasonable suspicion threshold
• timely human verification
• proportional duration of freeze

Relevance:

• Algorithmic AML systems cannot impose indefinite freezes without review
• Excessive reliance on automated suspicion creates liability risk

III. How Swiss Courts Treat Algorithmic Freezing Systems

Swiss courts do NOT treat algorithms as independent legal actors. Instead:

1. Algorithms are “tools,” not decision-makers

Responsibility always lies with:

• banks
• insurers
• fintech platforms

2. Strict liability applies for operational systems

Even if:

• AI misclassifies a transaction
• fraud detection system blocks legitimate payment
• reimbursement engine rejects valid claim

The institution remains liable.

3. Human oversight is legally expected

A completely automated freeze system without review is legally risky under Swiss standards.

IV. Typical Liability Outcomes in Switzerland

If an algorithm wrongly freezes reimbursement:

The institution may owe:

• Damages under contract law (Art. 97 CO)
• Repayment of funds (Art. 62 CO)
• Interest on delayed funds
• Potential regulatory sanctions (FINMA oversight)

Defenses generally NOT accepted:

• “The algorithm decided”
• “System flagged fraud automatically”
• “AI error was unavoidable”

V. Key Legal Insight

Switzerland follows a strict human accountability model:

Automation may assist decision-making, but it does not replace legal responsibility.

This makes Swiss law more conservative compared to jurisdictions experimenting with algorithmic liability theories.

VI. Conclusion

“Algorithmic reimbursement freeze liability” in Switzerland is governed not by a separate doctrine, but by traditional liability principles applied to modern automated systems.

Across Swiss Federal Supreme Court jurisprudence, the consistent rule is:

If an automated system freezes or denies reimbursement wrongly, the institution is fully liable under contract, tort, and unjust enrichment principles.