News / AI 'no silver bullet' to deal with potential chaos after de minimis ends

Those hoping AI will deal with the mass of declarations set to flood US customs offices from 2027, following the Trump administration’s decision to end the de minimis exemption on a global scale, are set for a rude awakening if they fail to properly recruit.

Access to the exemption has already been terminated for goods from China and Hong Kong, since 2 May, with warnings that customs transactions are only set to surge further.

Airforwarders Association (AfA) executive director Brandon Fried said “the sheer volume” of new transactions prompted by the 2027 de minimis discontinuation would “necessitate a greater human presence, to review, verify, and process these entries effectively”.

Mr Fried’s comments followed suggestions the administration had pinned its hopes on AI handling the additional workload, but there is little sign the technology will support training programmes.

“AI can certainly augment customs officers’ capabilities and make their work more efficient by flagging anomalies and automating routine tasks, but it cannot replace nuanced judgment, investigative skills, and physical inspections by trained officers,” Mr Fried told The Loadstar.

“We need both smart technology and a robust, well-staffed human element to navigate this new landscape effectively.”

If the US is to sufficiently staff the demand the 2027 change will provoke, it needs to begin training now, as the average customs agent requires two years to attain full certification.

Worryingly, the US is already understaffed in customs agents. An Oxford Economics report puts the shortfall at 5,000 and estimates that 27,000 will need to be trained to meet demand.

“Our confidence in AI’s capabilities as an enabler, allowing officers to handle more complex cases and larger volumes is high, but it cannot fully compensate for the increased manpower requirements needed to facilitate efficient and timely flow of goods,” Mr Fried added.

“AI, while powerful, is not a silver bullet or a substitute for the need to have more customs officers at our border entry points.”

Others are equally concerned, one expert on the issue telling The Loadstar that while “AI can enable people to process declarations faster”, it would struggle with “the peculiarities and individuality of declarations”.

Noting that the significant increase in workload for Customs and Border Protection (CBP) will, by extension, increase AfA member workloads, Mr Fried is keen to see a solution reached.

Scott Curtiss, head of field presales, manufacturing, at retail data provider Relex Solutions, told The Loadstar businesses could themselves benefit from AI as a means to tackle the potential chaos from the de minimis termination.

“For example, a retailer using advanced planning systems can simulate the impact of customs duties on products sourced internationally versus domestically,” said Mr Curtiss.

“Organisations can use scenario planning to build a playbook of potential responses – changing sourcing or adjusting pricing strategies – ahead of disruptions to allow them to execute strategy changes to maintain margins faster and less reactively.”