A rapidly expanding source of advice and search tool for knowledge, AI is an increasingly valuable tool in the workplace…
But how does it rate its own reliability when it comes to critical import and export compliance tasks, such as tariff classification?
Linda Middleton-Jones turns the eyes of AI on itself to answer a question of growing importance for those involved in global trade administration.
Over the years, we have featured the complex and daunting topic of Tariff Codes in articles, webinars, and training. Indeed, many of our specialists’ consulting engagements include some degree of work on classification. The result of their trusted and experienced advice has often been a combination of improved revenue, but most importantly, better compliance with regulations and laws that could otherwise prove costly.
I turned to Claude.ai with the prompt; “Would you recommend using AI to classify goods?“, and here’s what I got:
There are several important reasons why AI should not be solely relied upon for verifying commodity codes:
- Classification Complexity and Nuance
- Commodity classification often requires deep understanding of product composition, manufacturing processes, and intended use that goes beyond what’s described in product documentation. AI systems may miss subtle distinctions between similar products that have different tariff classifications, leading to incorrect code assignments.
- Regulatory and Legal Consequences
- Incorrect commodity codes can result in serious legal and financial penalties, including customs violations, duty miscalculations, and trade compliance issues. The stakes are too high to rely on AI systems that may have accuracy limitations or make errors that human experts would catch.
- Constantly Evolving Classifications
- Tariff schedules and commodity codes change frequently due to new trade agreements, regulatory updates, and emerging product categories. AI systems may not be updated quickly enough to reflect these changes, leading to outdated classifications.
- Context-Dependent Decisions
- Many classification decisions require understanding the specific context of how a product will be used, its market positioning, or regulatory requirements in different jurisdictions. This kind of contextual reasoning is challenging for current AI systems to handle reliably.
- Lack of Accountability
- When AI makes classification errors, it can be difficult to trace the reasoning or hold anyone accountable for the mistake. Human experts can explain their decision-making process and take responsibility for their classifications.
- Need for Human Expertise
- Experienced customs professionals understand the practical implications of different classifications and can navigate grey areas where multiple codes might technically apply. They can also identify when additional documentation or clarification is needed.
While AI can be a useful tool to assist and speed up the initial screening process, final verification should always involve qualified human experts who can ensure accuracy and compliance with trade regulations.