Ghana’s introduction of an AI-powered customs valuation system, Pelican AI, by the Ghana Revenue Authority has sparked controversy among importers and customs agents.

The system, designed to automate the inspection and valuation of imported goods by cross-checking declarations against global data in real time, was introduced to curb revenue losses from undervaluation, misclassification, and smuggling.

However, stakeholders, particularly the Association of Customs House Agents Ghana, have raised concerns about its opaque implementation and lack of consultation.

ACHAG President Yaw Kyei stated that customs agents support modernization but require clarity on how the system operates, warning that the absence of transparency is causing confusion and uncertainty at the ports.

A major point of contention is the system’s apparent override of internationally accepted valuation standards under the World Trade Organization, particularly transaction-based pricing.

According to Kyei, Pelican AI assigns its own valuation regardless of invoice value, disregarding factors such as bulk discounts, negotiated prices, and supplier relationships.

Stakeholders reported cases where identical consignments from the same exporter received different valuations, while items within the same shipment were assigned inconsistent values.

Some importers have reportedly experienced duty increases of up to four or five times, with consignments previously valued at GH¢10,000 allegedly reassessed as high as $52,000, creating unpredictability that affects business planning, pricing, and competitiveness.

Calls for transparency have intensified, with Michael Okyere Baafi, Member of Parliament for New Juaben South, formally requesting detailed information from the GRA under the Right to Information Act, 2019 (Act 989).

Citing reports from the Daily Graphic that the system is generating millions of Ghana cedis daily, Baafi submitted nine questions seeking clarity on the companies managing the system, the selection process, engagement timeline, and operational details.

Industry stakeholders continue to demand broader consultation, comprehensive training, and a review of the implementation process, arguing that without transparency, the AI system risks undermining business confidence despite government assurances that no new charges have been introduced.