The question of what happens to the border between Northern Ireland and the Republic of Ireland as the UK leaves the European Union is one of the most taxing challenges affecting the Brexit negotiations. Artificial Intelligence (AI) is not something that magically produces an answer to every difficult question – however, it’s already in use elsewhere around the globe and offers innovative new techniques to address challenges facing border and port management should the United Kingdom exit the EU Customs Union.
In matters concerning Brexit there has been an unfortunate habit of obscuring facts with fiction. In this particular domain that is perhaps justifiable as much of the work of the authorities is by its nature sensitive and secret. An open discussion of operational techniques and technical capabilities is hardly appropriate and is not furnished here.
The critical question is: without a physical border, can the UK maintain its constitutional and border integrity whilst honouring the Good Friday Agreement as a country outside of the EU customs union?
A revised approach and innovative thinking may show the way. One very workable scheme that could help to deliver expeditious yet secure operations at the national boundary comes in the form of an AI-based trusted trader scheme.
A scheme such as this would typically allow businesses to apply for pre-vetted status. Participating companies agree to supply sufficient information about their business to enable machine assisted decisions to be taken to vet each transaction. This means that using these schemes, vehicles do not normally stop for inspection at borders. This solution is being deployed in the Asia-Pacific region. Here, statistical and network analysis across a variety of data give early warnings to those responsible for the border. For example, it may be suspected that a business has changed ownership. Its risk profile would change if it passed into the hands or influence of known or suspicious persons. In such instances arrangements can be made to assess and check goods elsewhere along its route, at a suitable waypoint or delivery location or perhaps for surveillance to be initiated. In such a case the present close collaboration between the RoI and UK authorities would enable joint decisions to be made as to the approach.
The Internet of Things (IoT) offers another way to both expedite and secure a trusted-trader programme. Using on-board tracking systems that are fairly standard these days, vehicle journeys can be monitored for any dubious activity, such as detours and unscheduled stoppages. AI-systems may be used to develop activity signatures from data feeds from tracking devices to detect anomalous or potentially high-risk events. Bona fide truckers would clearly find the reduction in potential delays far outweighs any possible privacy concern and furthermore agreed data retention processes would be applied. These systems would alert officials to take near real-time action to mitigate risks for vehicles so-equipped and allow other techniques to focus on the less high tech.
How else does AI enable a soft border?
- Detecting immigration variances – AI can process flight, ferry and vehicle travel pattern information flagging behavior that is out of the ordinary
- Identifying complex evolving threats – working quickly enough (real-time) to cope with evolving threats. Risk assessment, using machine learning, is much more flexible than a rules-based system because it relies on what the data is saying. Rather than using a pre-set list of rules determined by humans in the first instance, it continues to learn and the models adapt accordingly.
- Reducing risk of increased border attacks – from existing and new opportunity criminals seeking to exploit the new arrangements.
The brilliance of using advanced analytics with AI and machine learning techniques is that insights can be extracted from data much faster and more cost effectively. In addition, authorities can make more informed decisions because far bigger heterogeneous data sets can be combined to deliver much deeper insight. For example, telemetry, transactional and other kinds of structured data can be combined with video, voice and other forms of unstructured data to uncover more valuable answers to real-time questions. More advanced use of AI would augment the commanders own experience and give guidance as to the most successful interventions available based on present resource deployment and historical records of similar cases.
Working in the real-time world, answering challenging issues around instant detection of fraud risk is what advanced analytics is designed for. Not to mention uncovering complex hidden networks of threat actors, including criminal entities, at scale and speed.
This can be achieved now in the UK and internationally, Intelligence-led border-control is becoming the norm right across the globe undaunted by curious UK domestic and foreign policy activities.
To find out more visit our public sector hub.