ICC Chair Harsh Pati Singhania Urges AI Governance and Digital Trade Rules

Introduction

Harsh Pati Singhania, Chairman of the International Chamber of Commerce (ICC), has called for the development of AI governance frameworks and digital trade rules to keep pace with the rapid digitisation of global commerce. This guide examines the compliance implications of AI-driven trade finance under UCP 600 and ISBP 745, identifies failure modes in AI-assisted document examination, and maps resolution pathways for banks and corporates adopting artificial intelligence in trade finance operations.

Failure Modes

1. AI misclassification of document discrepancies: AI systems may classify compliant documents as discrepant, or vice versa, due to errors in training data or algorithm design. This can lead to unjustified refusals or the acceptance of non-compliant presentations.

2. Lack of explainability in AI decisions: If an AI system determines that a document is discrepant but cannot explain the basis for its determination, the bank may be unable to justify its refusal to the beneficiary, exposing it to legal challenge.

3. Data privacy and confidentiality concerns: AI systems trained on trade finance documents may inadvertently expose confidential business information. Banks must ensure that AI training and operation comply with data protection regulations.

4. Over-reliance on AI without human oversight: Banks that fully automate document examination without human review may miss contextual nuances that AI systems cannot detect, such as the commercial intent of the parties or customary trade practices.

5. Regulatory uncertainty about AI in financial services: Regulators in many jurisdictions have not yet issued clear guidance on the use of AI in trade finance. Banks adopting AI may face regulatory scrutiny if their systems produce errors.

Resolution

1. Banks should implement AI as an assistive tool, not a replacement: AI should be used to flag potential discrepancies for human review, not to make final compliance determinations. This preserves the human judgment that UCP 600 and ISBP 745 contemplate.

2. Maintain audit trails for AI decisions: Banks should ensure that all AI-assisted document examinations are logged, including the AI's determination, the human examiner's review, and the final compliance decision.

3. Adopt explainable AI principles: Banks should select AI systems that can provide clear explanations for their determinations, enabling human examiners to understand and validate the AI's reasoning.

4. Engage with regulators proactively: Banks should engage with their regulators to discuss AI adoption plans and seek guidance on acceptable practices for AI-assisted document examination.

5. Develop internal AI governance frameworks: Banks should establish internal governance frameworks for AI use in trade finance, including risk assessment, testing, monitoring, and accountability procedures.

6. Train staff on AI-assisted examination: Staff who work with AI-assisted examination systems must be trained on the system's capabilities and limitations, including when to override AI recommendations.

7. Participate in ICC and industry working groups on AI: The ICC and other industry bodies are developing guidance on AI use in trade finance. Banks should participate in these efforts to shape emerging standards.

Conclusion

ICC Chairman Singhania's call for AI governance and digital trade rules reflects the growing importance of artificial intelligence in trade finance. While AI offers significant potential for improving efficiency and reducing errors, banks must implement it within a governance framework that preserves human judgment, maintains audit trails, and complies with emerging regulatory expectations. The integration of AI into trade finance document examination should proceed cautiously, with human oversight at every stage.

Frequently Asked Questions

Q1: Can an AI system legally perform UCP 600 document examination?
A1: UCP 600 does not explicitly prohibit AI-assisted examination, but it contemplates examination by a "nominated bank, confirming bank, or issuing bank" — which implies human involvement. AI should be used as an assistive tool with human final review.

Q2: What are the regulatory risks of using AI in trade finance?
A2: Regulatory risks include potential liability for AI errors, data privacy violations, and failure to meet the standard of care expected of banks. Banks should engage regulators proactively and maintain human oversight.

Q3: How can banks ensure AI determinations are auditable?
A3: Banks should implement logging systems that record all AI inputs, outputs, and human review decisions. These logs should be retained in accordance with the bank's document retention policies.

Q4: What role does the ICC play in AI governance for trade finance?
A4: The ICC is developing guidance on AI use in trade finance, including best practices for document examination, data governance, and accountability. Banks should participate in ICC working groups to stay informed.

Q5: Will AI replace human trade finance examiners?
A5: AI is unlikely to fully replace human examiners in the near term. The complexity of trade finance documents, the need for contextual judgment, and regulatory requirements for human oversight all favour a hybrid approach.

Source Notes

• "ICC to prioritise AI governance, push for global digitisation of trade documents: Singhania — The Times of India." Source context only; guide written from original analysis.
• "Next trade order needs stronger developing-world voice: ICC Chair — Business Standard." Source context only.
• "AI and data are the new raw materials for trade: ICC chief — BusinessLine." Source context only.
• "ICC will amplify Global South's voice in shaping global trade rules — CNBC TV18." Source context only.
• "ICC chief calls for digitisation of trade papers for ease of access — The Economic Times." Source context only.