Artificial Intelligence (AI) has the potential to revolutionise border management, streamline trade operations, enhance efficiency, and strengthen security when fully adopted by nations, experts say. Customs clearance processes, historically burdened by inefficiencies such as excessive paperwork, unpredictable delays, and costly errors, can be transformed by technology, according to David Smason, Chief Executive Officer of Cargo Seer AI, an AI powered cargo inspection platform. ALSO READ: Experts on how digitalisation can drive cross-border trade, expand African exports Studies show that delays at African customs average 12 days in Sub-Saharan Africa, compared with seven and five-and-a-half days in Latin America and Central and East Asia, respectively. Speaking at the Africa Trade Development Forum on December 3, Smason highlighted that AI has the potential to streamline procedures and improve efficiency at customs between countries. “With AI, customs can analyse shipping data in real time, flag anomalies almost instantly, and even predict trade patterns to minimise congestion at ports,” he said. Smason also stated that beyond speed, AI offers enhanced security. “Advanced systems can spot suspicious patterns hidden in trade data, such as misdeclared goods, unusual shipping routes, or red flags in customer histories, swiftly addressing customs delays,” he explained. ALSO READ: How digitalisation is transforming trade in Africa For instance, he added, machine learning algorithms can identify counterfeit products, illegal wildlife shipments, or smuggled items before they ever reach consumers, something that otherwise could escape human observation in time. “Such innovations are not only safeguarding trade but also protecting the environment and public health by identifying risks hidden within the global supply chain, boosting trade, and reducing delays at customs,” he noted. Potential to bring efficiency While some worry that automation could replace jobs, many customs organisations see AI as a collaborative tool rather than a competitor. Technology, they argue, takes care of repetitive tasks, freeing up human agents to focus on complex decision-making. According to Global Freight Insights, a logistics market research firm, companies using AI-enabled customs software have reported up to a 30 per cent reduction in clearance times. Albert Atambo, Chief Manager of the Business Transformation Office at the Kenya Revenue Authority, said that what a human analyst takes five to 30 minutes to analyse, AI is capable of completing in just a few seconds, saving significant time. ALSO READ: Digital technologies could unlock Africa's trade potential, says Ngirente “By integrating scanners between Kenya and Uganda, we have seen an immense increase in image data, nearly doubling the workload for our analysts. AI addresses this challenge by analysing images instantly, drastically reducing delays at ports and borders and tackling bottlenecks caused by manual processing,” he observed. But Atambo indicated that AI does not only perform image analysis, it can also analyse consignments, combining data from customs systems, domestic taxes, and partner states to make informed decisions. “By comparing scanned images from Mombasa to those at the Kenya-Uganda border, it ensures transparency and consistency in cross-border transit, promoting trust and economic integration within the East African Community,” he said. Atambo also observed that the goal is to use AI to unify all technologies, including smart gates, electronic road parking systems, and scanner data, into one powerful system that ensures efficiency and accuracy. Logistics experts like him argue that solving the current challenges in image analysis and cross-border trade would be nearly impossible without AI, thereby limiting the pace of modernising customs management.” ‘AI could simplify taxation processes’ According to Stella Kwera, Manager of Data Science at the Uganda Revenue Authority, AI has potential to simplify the process for taxpayers, such as providing quick and accurate answers to common inquiries, including how to file and how much to pay, saving time for both taxpayers and authorities. “By utilising generative machine learning models, we can offer these answers more efficiently. Additionally, AI allows us to identify potentially fraudulent taxpayers by analysing patterns in their data, which can ultimately boost the country's economy,” she said. Generative machine learning models are a type of AI that use machine learning to create new data that resembles the data they were trained on. These models can generate new images, text, audio, and video. Flavia Busingye, Director of Customs at the East African Community (EAC) Secretariat, highlighted the need for enhanced collaboration among EAC member countries and other nations to fully embrace AI in customs processes. “When I reviewed our digital strategies and protocols, I realised there are no specific or standalone provisions relating to AI. This calls for further engagement to adopt AI with clear strategies that enable our customs administrations to leverage available technologies,” she said. “As data security and trust are critical components in regional AI collaboration, we must think about better data protection and security. There is often fear around sharing data, so it is vital to establish measures to preserve and protect each nation's data,” she added. Busingye also underlined the need to build capacity for AI use across the region, highlighting that some countries are at more advanced stages. “But as a continent, we need to operationalise these technologies collectively and ensure inclusivity.” Need for ethical standards Felicien Mwumvaneza, Commissioner for Customs at the Rwanda Revenue Authority (RRA), highlighted the ethical and legal considerations of employing AI in risk assessment and predictive inspections, stressing the importance of fairness and minimising disruptions in data collection during implementation. “The cargo trucking system, for us, has been truly phenomenal. It has significantly reduced risks, mitigated corruption, and minimised costs in our operations,” he explained. “From the solutions we have tested, one issue that consistently arises is data accuracy. Whatever solutions we adopt must be context-specific and customised to the trade practices of our region,” he added. Mwumvaneza explained that the customs system in the region is very detailed. It includes nearly 6,000 specific categories to classify goods, and some of these categories are broken down into over 500 smaller, more specific classifications. This helps ensure accurate tariffs and proper tracking of goods, but it also makes the system complex to navigate. “This complexity requires that any AI system we adopt must align with our unique needs. The legal battles should underscore the need to carefully consider privacy and copyright laws when deploying AI,” he said, A lack of comprehensive regulatory framework, he maintained, leaves countries to navigate uncharted waters.
