Round-up of the 2017 WCO Technology & Innovation Forum
19th March 2018
From 31 October to 2 November 2017, for the sixth time, the WCO organized its Technology & Innovation Forum, a platform for leading technology experts and decision-makers from various border agencies, as well as representatives of the academic world and international organizations, to discuss technologies and how they can add to the capabilities of Customs officers as they work to address border challenges.
This article highlights some of the discussions at the event that was held in Tokyo, Japan, and which focused on the pros and cons of new technologies, the impact of disruptive technologies on borders, the trust between humans and machines, the use of geospatial data in the Customs environment, and the utilization of data analytics – a recurrent but important issue which the WCO chose to highlight as its theme for 2017.
Just a few years ago, “blockchain” was seen as a technology, use of which was limited to Bitcoin or other cryptocurrencies. Today, everybody is thinking about applying this revolutionary technology that can be used for non-currency purposes too. For example, some companies and organizations are already trying it in areas relating to supply chain record-keeping and the “track-and-trace” of products, underscoring its huge potential.
Blockchains can store a range of records including payment transactions, sales records, purchase history, corporate accounts and retail pricing history as well as future changes in pricing. It can also record non-transactional data such as title records, trademark and patent information as well as travel logs to name a few. In essence, blockchain is a continuously expanding database that stores transactions in real-time.
Although blockchain is being talked about a lot, it is not a one-size-fits-all solution, experts highlighted. Rather than looking at the solutions that exist and considering where they can apply in an organization, one should analyse the problem that one is trying to resolve and see whether blockchain can add value. The technology is most valuable when several parties are connected and need to access the same data, but do not necessarily trust each other, and when there is no trusted third party mandated to collect and share the data.
How to keep up with changes and cope with them is one of the questions which Forum participants raised. The fast adoption of blockchain technology demonstrates how fast the pace of innovation really is. By the time one makes a decision to buy something, there is already something else on the market.
Speakers pointed out that it is critical to understand what is coming next, and to always keep aware of developments by looking, for example, at how other services or businesses are working. Solutions need to be piloted and new profiles and talent must be hired. In order not to be stuck with a specific solution, as it might change or be replaced by better performing technologies, some countries have been investing in information technology platforms that allow them to more easily change from one technology to another without disruption.
As the increased adoption of technology is fast-changing the role of Customs officers and the way that they perform their duties, it is also important for leaders of Customs administrations to communicate effectively and regularly, in order to instil a change and innovation mind-set. The need for Customs officers to understand the impact of technology on their organization is, therefore, imperative.
Customs officers also need to trust the technology that they are using. Today, it is quite natural to trust machines such as licence plate readers, passport readers, and chemical detector devices. However, when it comes to machines dotted with artificial intelligence (AI), things become more complex.
How much of it should be placed in “smart” machines? And how can the creators of these machines help to answer this question? According to them, the threshold has been crossed, as a number of initiatives already show that machines can be trusted with making decisions. This can be seen in the field of healthcare, for example, where machines monitor parameters, predict issues and even generate correlations, patterns and hypotheses to identify personalized therapy options for patients. Another good example is in the field of image recognition, where the accuracy of machines far exceeds that of humans.
To find indicators of known and unknown risks, traditional methods of analysis are slowly being replaced or complemented by cognitive computing solutions, which include data mining, machine learning, and natural language processing capabilities. Cognitive computing can help Customs officers make better decisions in a wide range of different scenarios by understanding vast amounts of unstructured data that is inaccessible to conventional IT systems. As long as the machine can explain its reasoning and how it generates its hypotheses, there should not be any trust issues, agreed the panellists. Moreover, the final decision to control or inspect a good or a traveller is not in the hands of the machine. Machines should enable one to get a smaller subset of transactions for experts to review. They also enable one to dispense with manual tasks and some of the analytical steps.
Various non-intrusive inspection (NII) suppliers explained that their automatic detection tools have, during pilots, proved to work well: algorithms enable machines to recognize objects and to distinguish between commodities. However, building such functionalities requires access to a lot of data and there are still very few solutions that have been implemented which one can learn from. There is even less information on the performance of the systems in real situations, although being able to measure results in terms of facilitation, revenue, and security is critical to build trust and to know whether one is going in the right direction.
The fact that we trust smart machines to make a decision raises another question: who should be accountable for the decision that the machine makes? For the moment, accountability is fragmented among Customs officers. If smart machines make the decisions, this will weigh heavily on the shoulders of managers.
Data and data analytics
Although all participants agreed on the need to establish data at the core of Customs activities and to transform Customs administrations into data driven organizations, countries are at different stages when it comes to turning this ambition into a reality.
“Our vision of the future is that data will be collected at all stages of the border continuum and analysed using analytical tools, the results of which will be submitted to an officer,” explained a speaker.
A panellist stated that his administration, having to manage difficult borders in areas where terrorist groups hide out, is investing in all sorts of equipment, including drones to get real-time data on the movement of people, as well as in a new information management system that will enable it to collect and manage more data.
Another panellist pointed out that in his country, which has to manage “fragile borders,” there is a lack of IT infrastructure to collect certain data, and that officers, especially those working at the borders, are not prepared for such a process and do not understand the importance of data analysis and what is required from them. Any adoption of technology would require training, and should take into account human capacity.
There is an area where the electronic collection of data is lagging behind: the processing of international postal items continues to be a very manual driven process. On this topic, participants heard about the implementation of an electronic data exchange pilot project between Australia and New Zealand’s Post and Customs administrations (the pilot project is presented on page 27 of this edition of the magazine).
The benefit of using corporate data in the fight against fraud was also discussed. The average Customs administration’s ability to detect and prevent fraud is not solid enough, explained a panellist. This was due to the proliferation of companies and the ease through which they can be established, in conjunction with the ability to mask criminal activities behind complex international operations and global structures. Criminals use corporate entities as easily as they use individuals. It is not uncommon to find companies with an ownership structure spanning 23 countries, and having more than 150 companies under its structure, with at least 50 individual shareholders.
One of the ways to fight fraud, or crime is to look at undeclared relationships, where undervaluation or false declarations can thrive. However, obtaining accurate data on private companies is a challenge. Corporate ownership structures are very often complex and evolving, and one needs to verify, “clean” and guarantee the quality of the data. The data will also have to be standardized to enable systematic data analytics. In addition, unique identifiers for entities will have to be created to accurately link them in order to create holistic networks. Customs administrations may also want to rely on third party providers to be able to trace the ownership of entities and analyse links between these entities and the people controlling them.
Collected external data is integrated into an internal data platform, and the enriched, combined dataset is then ready to be ingested for risk assessment purposes. If there is too much corporate data for analysts to review without the help of technology, having the data and the technology to analyse it alone is not enough given the level of complexity of fraud. Relying on human analysts is critical to remove false positive results.
Forum participants also discussed the added value of mapping data and the use of geospatial technologies. Maps are a useful tool in simplifying complex information for the purpose of educating groups, persuading people, or understanding ideas or situations. Maps are visuals, and visuals are compelling. It has been said that 90% of information transmitted to the brain is visual, and visuals are processed 60,000 times faster in the brain than text.
Maps are not only a great way to report information that is known, they can help one discover what is not known. They can also assist one in checking whether something that one is not sure about, but feel strongly about, is correct. Last but not least, mapping data offers a great opportunity in terms of cooperation, as it enables one to open a dialogue with experts from different backgrounds and even different countries, each drawing his/her own analysis according to his/her own perspectives.
Several cases where the addition of a spatial component to data enabled light to be shone on problems or conclusions to be drawn quickly were given. The WCO explained how it used geospatial data in a research programme related to the Lake Chad region, the lake being at an intersection between several countries, a trade hub for Sudan, Central Africa and East Africa. The border was defined as a large buffer zone, an area structured by the economy of the border where Customs would apply its controls. The objective was to look at what happens between the entry points at the border, what were the dynamics of trade in goods and flows of people, which roads were used and how frequently, etc., and to provide officers on the ground with relevant information they could act upon.
In such a region, as the Lake Chad one, it is crucial to know where to allocate resources as any deployment requires huge resources. It is also crucial to understand and make apparent the impact of the security measures on the economy of these border zones to inform policy makers. Other projects which were presented to participants included the analysis of truck traffic, informal trade flows, and activity at ports.
Customs collect very little geodata, a collective term for all kinds of data with a spatial reference. Ideally, one should have data from the field and from communities living in the border areas as well as data from transporters crossing such areas. But, one can build maps from data that is available publicly. There is a lot of georeferenced data available on the geography of a territory (elevation, waterways, vegetation), on infrastructure (roads, paths, borders), on demography and settlements, and on an economy, including market and infrastructure location. Satellite images are also easily accessible, and news agencies are very good sources of information.
A variety of software is available to capture, store, manipulate, analyse, manage, and present spatial or geographic data, known as a geographic information system (GIS). Some are open-source software, others are traditional proprietary software. Forum participants were presented with different technologies, from map making tools to very sophisticated geospatial knowledge management solutions which connect real-time information on the positions of items, data from connected objects (Internet of Things), data from satellites and CCTV, and data from remote sensors.
If data analysts are needed more and more, the work of chemists is also critical to law enforcement agencies, for example, to determine the origin of seized illicit substances. During the Forum, Japan Customs presented its methamphetamine signature programme that was inaugurated in 2008, and the tools it has at its disposal to analyse the drug. By analysing the chemical profile of seized drugs and the precursors used to make them, the administration is able to identify the origin of the products per region, what routes they travelled, and compare cases in order to identify drug networks.
Remote and centralized screening
The benefits of integrating and centralizing radioscopic images generated by NII equipment were once again highlighted. Solutions providing for a common interface have been designed that enable all the images produced by the machines to be seen on the same screen, using the same analytical tool. Their benefits include the possibility of centralizing the management of scanning operations, and of handling inspections from a remote control centre. Projects are underway to integrate the inspection results of other detection systems as well as information sources of different law enforcement agencies to get a complete and accurate security profile from a common viewer workstation.
Participants were also updated on the evolution of the “unified file format” (UFF), the international standard for scanned images and associated metadata produced by NII equipment, which is being developed by the WCO and NII suppliers. The project will be completed in the coming months.
Today, in many administrations, officers are mobile and can work remotely as they have access to applications via tablets or smart phones. For example, inspections can be held on site from A to Z; travellers can be checked with a phone with a fingerprint reader attached to it. Most tools or databases should be available through mobile applications, agreed the speakers.
Voice recognition software has become so exceptional that when it comes to the future of this technology, the consensus seems to be voice recognition in everything. With this in mind, it would not be far-fetched to assume that keyboards and similar control panels will slowly be phased out of all devices as they gain the ability to simply listen and act on human voice commands.
Facial recognition software is based on the ability to differentiate between a face and the rest of the background, measure the various features of the face, and compare it to a database of stored images. Many administrations have already invested in the tool or plan to do so, to more easily identify passengers to be controlled at airports and to cut down on crimes such as identity fraud – Japan, who will be hosting the Olympics Game in 2020, is a good example. Participants who were given the opportunity to test such a tool at the Forum were very impressed.
It would be impossible to exhaustively review all the technologies presented during the Forum. Devices for the detection and identification of products were displayed at the exhibition, as well as handheld X-ray instruments. Participants were also able to watch videos and presentations on the latest generation cargo scanners, of container gas testing and ventilation solutions that protect the health and safety of unpacking staff, and much more.
Once again, the event proved to be a useful tool to start the conversation between Customs and technology provider representatives, and hopefully move the implementation of technologies forward.