The Algorithmic Gavel: Analysing automated customs valuation, ICUMS and the rule of law in international trade

The Ghana Revenue Authority (GRA) has implemented the Integrated Customs Management System (ICUMS), an automated customs valuation system, to modernize customs operations in Ghana. This digital platform aims to enhance trade facilitation, revenue collection, and administrative efficiency. However, the reliance on ICUMS, which incorporates artificial intelligence tools like the Publican AI vetting system, has sparked legal and economic concerns regarding the role of algorithmic decision-making in customs administration.

Under the WTO Valuation Agreement, Ghana has legal requirements for customs valuation that must be strictly followed. The primary basis for assessing customs duties is the Transaction Value, defined as the actual price paid or payable for goods sold for export to the country of importation. Customs authorities are strictly prohibited from skipping any step in the six-step valuation hierarchy established under Section 67 of the Customs Act, 2015 (Act 891).

They cannot rely on artificial values generated by AI systems unless they have thoroughly investigated and disproven the use of a more appropriate valuation method.

The WTO Valuation Agreement allows customs administrations to discard the transaction value only under strict statutory exceptions, such as fraudulent transactions lacking arm's length pricing or the absence of identical goods in recent imports. However, ICUMS's use of pre-programmed valuation data may override genuine transaction values, potentially compromising compliance with Ghana's tax laws and WTO obligations.

This practice raises important questions about the rule of law, due process, and the proper role of algorithmic decision-making in international trade.