The digital transformation of internal security in the EU, AI and the role of eu-LISA

​By Krum Garkov, Executive Director of eu-LISA

The ongoing digital transformation of internal security is a fact – today efficient internal security means information driven internal security. In addition, the COVID crisis only accelerated the digital transformation of traditional business models and internal security is no exception to this trend. Furthermore, if digital technologies are the engine of modern internal security, then the fuel of that engine is data. This is why utilizing big data for the purposes of internal security is a new paradigm to achieve better efficiency in the work of law enforcement officers.

Better efficiency is necessary because, in the long, run the nature and range of security threats will only grow. At the same time, it is unlikely that in the coming years there will be a substantial increase of the workforce engaged in the area of internal security.

Utilizing big data only though, will not be enough. Data is just the raw material. What really matters is the information created from it and put in the hands of practitioners, enabling them to take better and informed decisions.

However, the volumes of data and information are constantly growing and that trend will continue in the future. To avoid being flooded with raw data and information, law enforcement practitioners should be provided with a better tool set to handle it. This is where the role of the Artificial Intelligence (AI) solutions come in.

AI is one of the most promising disruptive technologies that could create tangible benefits in the area of internal security. It could enable law enforcement practitioners to follow an integrated and risk-based approach to address present and future security threats. Hence, the question is not whether to adopt AI but when and how…

In the EU, the great potential of AI for internal security has been recognised and has become a political and operational priority, reflected in the EU Security Union Strategy.

eu-LISA has also recognised the potential of AI and is pro-actively looking for ways to implement it practically in the areas of the Agency's responsibilities as per its mandate and in its day-to-day operations. At the same time, the Agency does not look at AI in isolation, but as an essential building block of the new information architecture for internal security and border management, currently under implementation by eu-LISA.

In July 2020, the Agency published its research report 'Artificial Intelligence in the Operational Management of Large-scale IT Systems'. The report provides insight on the way eu-LISA can explore the potential of AI in its internal operations and for benefit of its stakeholders. 

Furthermore, eu-LISA contributed to several EU-wide studies that explored different aspects of the potential use of AI in the JHA domain, in particular:

  • A study on opportunities for AI in border control, coordinated by DG HOME;

  • An impact assessment on hosting a data space for Law Enforcement. 

In the context of the new information architecture for internal security and border management, eu-LISA identified several use-cases for implementation of AI-based solutions in the coming years:

  • AI for ETIAS, to enable automatic screening of the travellers on the basis of pre-defined risk indicators and screening rules;

  • AI for CRRS, to increase the analytical tool set provided to the Agency's stakeholders;

  • AI for sBMS, to improve and enhance the accuracy of biometric matching algorithms;

  • AI in the EU e-Visa Proof-of-Concept, presently under development at eu-LISA. 

eu-LISA is considering launching two additional Proofs of Concepts to explore the potential of AI in its internal operations:

  • For the use of a virtual assistant/chat bot in its Service Desk; 

  • For the implementation of an AI-based solution for predictive analytics of IT infrastructure and/or network with the aim to improve availability of infrastructure and reduce failures and downtime.

Last but not least, eu-LISA intends to establish a horizontal Working Group on AI, to further explore the practical implementation of the findings of DG HOME's study 'Opportunities and challenges for the use of artificial intelligence in border control, migration and security'. This would be done in the context of the systems entrusted to the Agency and the new information architecture for internal security and border management.

At the end, I should note that while the implementation of AI solutions might look quite technical, in reality the main challenges are not technical at all. To me, the four main challenges that have to be recognised and addressed in order for the implementation of AI solutions to make sense and to provide tangible benefits, are:

  1. Political and operational leaders need to recognise the potential of AI and to promote and steer the implementation of AI solutions;

  2. A new balance between use of big amounts of data and data protection/privacy safeguards needs to be found. EU legislation needs to evolve and establish a framework for an ethical use of AI. At the same time, it has to match the evolution of technologies and the needs of practitioners.

  3. Substantial effort is required for training and capacity building to enable law enforcement practitioners to explore the benefits of AI fully.

  4. Trust needs to be built in society that the capabilities of AI will bring an important added-value and will be used for legitimate purposes.

If these challenges are not recognised and addressed properly, the technical excellence of the solutions put in place will bring limited benefits.  

In eu-LISA we firmly believe that AI is an indispensable building block of the ongoing digital transformation and that its potential and capabilities should be explored without delay. The Agency is committed to this journey and prepared to play a pro-active role in it. At the same time, it is clear that the proper implementation of AI for internal security in the EU can only be achieved as a collective exercise and should be based on synergies and complementarities between all involved actors, and close collaboration between the public and private sectors.

Operational Management; Stakeholders; European Institutions; Member States