AI & legal complexity management

Compliance complexity is growing faster than our capacity to manage it manually, and this recording examines why artificial intelligence (AI) — despite its promise for logistics — remains a "blind spot" when it comes to customs and trade-compliance decisions, and what it would take to fix that.

Topics covered:

• Why e-commerce and low-value package growth (with error rates cited as high as 86% in a real customs check) is overwhelming manual compliance checking
• Geopolitical shifts, new tariffs, export controls and sanctions regimes increasing regulatory divergence and compliance cost between countries
• New due-diligence obligations (forced labour, deforestation, carbon border adjustment) adding data-collection burden on supply chains
• Where general AI tools already help today: extracting and validating shipment data, checking HS code/product description matches, and screening counterparties against sanctions lists and adverse media
• Why general-purpose large language models are unreliable for regulatory compliance, including a real example of an incomplete AI answer on shipping ball bearings, and cited error-rate research on specialised legal AI tools
• "Norm engineering" as a structured methodology for breaking legal text into consistent, source-traceable building blocks (who a rule addresses, what it requires, under what conditions) to reduce AI hallucination
• The Large Language Models for Logistics Management (LLM4LM) AI Lab, a Dutch-government-backed innovation lab combining norm engineering with logistics AI pilots, and the idea of a future "trust infrastructure" linking administrative and physical shipment data

For a broader overview of the topic, please watch the full recording. The slides are available in the Resources section.

Please note that this summary was generated using AI, based on the recording and available slides.