6 key considerations for adopting AI technologies for Customs
It is recognised that artificial intelligence (AI) represents huge opportunities for organisations like Customs to automate business processes and make their operations smarter. However, many Customs administrations are failing to adopt AI because they are still uncertain about how to approach it. These are key considerations in unlocking the benefit of AI.
1. Need for political commitment
Because AI adoption requires investment and disrupts the way Customs does business, Customs executives should act as champions and be responsible for validating the AI initiatives, approving the selection of the use cases that will be considered, mobilising resources, and enforcing decisions.
2. AI-savvy workforce empowerment
A successful AI adoption in any organisation will require the development of both technical and managerial capabilities.
The management of AI technology also involves new managerial skills such as judgement-oriented skills, creative thinking and experimentation, data analysis and interpretation, and in-depth domain expertise. For example, fraud targeting AI applications may reduce the time managers spend looking for risk patterns but increase the requirements for interpreting the outputs combined with their expertise and ethics and drawing the final decision.
AI requires new technical job categories such as the next generation of machine learning engineers and AI products managers at the technical level.
Customs administrations that have been best at adopting AI are using multiple talent acquisition paths: internal talent upskilling and reskilling, new talent hiring, management and promotion.
3. Building on solid digital foundations
AI works best when it has real-time access to large amounts of high-quality data. Therefore, it is critical to determine if the current IT systems and processes (information digitalisation and storage, paperless, accessibility, level of quality, computing capability and security) are sufficient to support the selected AI initiative adequately.
4. Personal data protection and privacy
AI requires access to important volumes of data. Customs policymakers need to carefully assess whether existing data access laws should be updated to maximise AI benefits. For example, when it comes to personal information, appropriate protection and privacy laws, data anonymisation requirements, and similar policies that balance privacy concerns against AI's benefits must be considered.
5. Integration of AI technology with existing legacy systems
In general, a legacy system that runs on heavily outdated software code does not work with modern Application Programming Interface (APIs), making it not possible to connect with the latest technologies.
Adopting AI technologies may imply updating existing components or rewriting part of the legacy code into a modern stack. This provides a new interface to the legacy system, making it easily accessible to/for the modern AI software components.
6. Black-box effect and usability
It is difficult for people to trust AI tools that make essential decisions without transparency about the rationale behind them. Because AI systems will assume responsibilities that used to be performed by humans, people must understand how these systems make decisions. The lack of transparency—the black-box effect—increases adoption resistance (Kafando et al., 2014).
Good AI solutions should have user-friendly and intuitive interfaces that allow Customs end-users (generally non-technical personnel) to understand the logic behind the complex AI algorithms to build their user experience quickly and efficiently capture the expected value.
Cotecna has been using artificial intelligence for years to power its risk management system – CRMS. For more support and information on implementing artificial intelligence technologies within the Customs environment, contact us at governments@cotecna.com.