compliance e AEO

AEO, customs compliance and artificial intelligence: the next frontier of the customs

Novembre 1, 2023 /

The aim of this article is to start to outline and analyze the impacts of the artificial intelligence (AI) on the customs compliance and AEO.

Firtsly, it could be interesting to understand the meaning of “artifical intelligence”.  Indeed, according to John McCarthy (emeritus professor at Stanford University in 1955) definition, artificial intelligence is “the science and engineering of making intelligent machines”. In practical terms, artificial intelligence allows the performance of actions normally done by human intelligence. By means of algorithms, the artificial intelligence system can:

  1. Learning;
  2. Reasoning;
  3. Improve its own creativity;
  4. Self check and control.

From the point of the view of the authorized economic operator (AEO), the development of  artifical intelligence approach should brings our attention of the following points.

The AEO is the unique authorization by which, the owner is recognized as reliable. This is the main peculiarity of this customs decision: indeed, this is the unique authorization which modifies the status of the economic operators  which, therefore, is, in the scheme of the customs obligation, the debtor of the customs duties and other taxes due for the import operation. Such a reliability is granted and mantained by the AEO only if this company is able to prove: a) very specific and high professional care (diligenza qualificata) in fulfilling the processes directly and indirectly linked to customs operations; b) the monitoring of all the main elements which can affect the customs/trade compliance.

The AI approach to the customs affairs requires, from a compliance point of view, that, the AEO company:

a) Is able to explain how AI supports the organization to meet this criterion laid down by article 39 UCC: “…the demonstration by the applicant of a high level of control of his or her operations and of the flow of goods, by means of a system of managing commercial and, where appropriate, transport records, which allows appropriate customs controls…”. The AI, indeed, should make better the controls, the management of the operations and harminize it with trasport records and needs coming from the customs regulations. In other words, it is expected that AI system is able to check the peculiarities goods imported/exported/manufactured and find out the proper HS code, the expected number of compliance checks and needed monitoring of troubles occured during the transport and customs operations. Thus, the applicant for an AEO or the onwer of this authorization should explain to customs the rationale of the AI systems.

In particular, in the questionnaire for self assessment, there are the following questions to be answered:

  • 3.1.1 “…Does your accounting system facilitate a full audit trail of your customs activities or tax relevant movement of goods or accounting entries? If yes, please describe the essential features of this audit trail…”;
  • 3.3 “internal control system”. This chapter of the SAQ is made of three questions about internal controls and audit. This is a section that should be very impacted by the AI; indeed, it could be interesting to see that the AEO checks, by means of periodical audits, how its own AI system works
  • 3.4  “flow of goods” sections and 3.5 “customs routine” could be some filed where the AI is or will be very useful.

b) the self assessment questionnaire requires the evidence of the practical standard of compentence or professional qualification. The applicant of an AEO or the owner of this status, should do prove that their AI system is able to “professionally and carefully” manage customs knowledge data according to the company needs? For us, it is more realistic to think that the responsible for the management of AI system must be prove their customs knowledge.

c) Last but not least, we should consider, in this short paper, another obligation of the AEO: according to article 39 letter a) of the UCC, the applicant of the AEO has to prove the “…the absence of any serious infringement or repeated infringements of customs legislation and taxation rules, including no record of serious criminal offences relating to the economic activity of the applicant….”. Among the list of “serious infringement” we should consider tha possible breach to the intellectual property rights performed by the AI. The EUIPO in the report “…Study on the impact of the artificial intelligence on the infrigement and enforcement of copyright and designs” lists the “…following types of IP are the main focus of the study:

  • copyright – digital content;
  • copyright – applied art;
  • copyright – other copyright-protected physical products;
  • rights related to copyright – digital content;
  • sui generis protection of databases – digital content;
  • registered Community designs – products;
  • registered Community designs – digital items;
  • unregistered Community designs – products;
  • unregistered Community designs – digital items;
  • registered design rights in the European Union Member States (EUMS) – products;
  • registered design rights in the EUMS – digital items…”.

d) Risks assessment: According to the guidelines on AEO released on 2016 “…The organisation of an economic operator can be a complex system involving many interrelated processes. An AEO should focus on processes, management of risk, internal controls and measures taken to reduce risks…”. It is clear that these tasks can be performed and boosted by means of the AI system. Indeed, the mentioned guidelines of the EU underline that “…The management system should allow for: – continual cycle of identifying needs or requirements, – evaluating the best means for complying with the requirements, – implementing a managed process for applying the selected management actions, – monitoring the performance of the system, – maintaining evidence of the application of processes used to meet business objectives, and identify functional or business improvement opportunities, including reporting mechanisms on gaps, incidental mistakes and possible structural errors…”.

From the Customs authority point of view, the AI can support:

  1. The improvement of effectiveness in executing internal core processes: e.g: risks assessment and customs duties collection;
  2. The promotion of the automation of routine and time-consuming tasks;
  3. An approach able to provide the state-of-the-art governance; it means check and control the processes;

In particular, WTO and WCO in the report “…The role of advanced technologies in cross-border trade: A customs perspective” 2022 listed, in a more precise way, the activities of customs authority that can improved by the AI: “…Post-clearance audits and controls, Automated targeting systems, Tariff misclassification and non-compliance with tariff advice, Misuse of concessions (including tariff concession orders, by-laws, free trade agreements and origin masking), Detection of anomalies in high revenue areas (including excise equivalent goods i.e. alcohol, tobacco, petroleum), Detection of dumping and countervailing anomalies, Detection of undervaluation and overvaluation, anomalies Refunds and drawbacks non-compliance Detection of prohibited goods; Monitoring service delivery performance in real time; Providing historical insights into customs statistical information for future planning and forecasting; Compliance risk scoring in commercial and trade activities; Identifying low value courier and postal shipments to improve risk assessment; Identifying low risk individuals at borders; Developing advanced analytics for AEOs…”.

A practical example is given by Carla Amaral de Andrade Junqueira, ‘Recent Customs Reforms in Brazil’, (2023), 18, Global Trade and Customs Journal, Issue 11, pp. 450-453 [abstract] “….Brazil has been actively addressing the challenges posed by the rapid growth of e-commerce within its customs framework, focusing on enhancing efficiency while maintaining control over trade flows. Two notable strategies employed by Brazil is the utilization of cutting-edge technology of artificial intelligence (AI) and a Trusted Trader Scheme, known as Authorized Economic Operator (AEO) program. This scheme acknowledges compliant and reliable economic operators, granting them a status that comes with various benefits, particularly in the realm of customs procedures. In addition, the application of automated risk analysis, often integrated with AI, has also revolutionized customs procedures in Brazil”.