Discussions focused on the digital transformation of the JHA area and the long-term vision for this process, the opportunities that technology and innovation are bringing to the internal security and border management domains, as well as on the implications of the COVID-19 pandemic on the acceleration of digitalisation in this area.
The benefits and possibilities for the law enforcement and border management communities stemming from the new information architecture and challenges of redesigning the Member States' business processes were also discussed. While embracing the novel solutions offered for improving internal security, the importance of ensuring the fundamental rights throughout this process was stressed.
The event was opened by European Commissioner for Home Affairs Ylva Johansson, who unfortunately could not attend the event live, but addressed participants in a video message:
Two panels tackled two very important questions: "Long journey from silos to interoperability – what lies ahead?" and "The new information architecture in JHA domain: endless possibilities?" and also posed a couple of flash polls to the audience:
One of the key takeaways from the event is that digitalisation in the JHA area has become more important than ever, as it contributes directly to resilience. In the current pandemic situation, it has become particularly evident how indispensable the ability to have fast access to information is; how important it is to ensure business continuity.
We should not stop here. We need to take advantage of modern technology, including interruptive ones, to facilitate and secure the life of European citizens. The Schengen Area needs to develop further, with a view to ensuring faster and more efficient border crossing. Artificial Intelligence will be a game-changer when processing vast amounts of data and looking for security gaps. The question is not "whether", but "when and how" to apply AI-based technologies.
The discussion focused on the challenges in using automated face recognition systems, as opposed to humans (e.g. border guards). While new technologies (such as deep learning) may outperform humans in face recognition, both humans and machines have problem with some common aspects.
Other points covered during the debate:
“There is only one way for quality input” was the conclusion of the first discussion panel: and that is only by ensuring it at the acquisition stage (i.e. when the photo is taken). Therefore, the more restrictions you impose in the acquisition context, the more you limit the variability of conditions, the better quality you obtain in the end. The quality of data used is also vital when it comes to AI - and specifically training AI-based systems.
The International Air Transport Association (IATA) presented their vision for an end-to-end biometric (touchless) passenger journey. A person's identification data is stored only on their electronic device (smartphone), and is shared only when needed with authorised parties, thus avoiding any GDPR concerns.
Natural Language Processing (NLP) technologies are especially useful in circumstances where biometrics of an individual are not available, or the use of biometric identification is not feasible, to ensure rapid and precise identification of persons on watch lists, for example.
Discussions and presentations touched upon privileged user access, potential breaches and the zero-trust policy to defend against such breaches.
Language itself can pose a number of challenges when it comes to biographic data: names of individuals, names of cities/ countries are pronounced differently in different languages; geographical changes can also be recorded differently, for instance birth location; transcription makes it even more complex. Therefore, a multi-cultural name matching solution could be considered.
In addressing data quality, we must consider three main aspects: organisation (e.g. data governance framework), data flows and data quality management architecture supporting data quality processes.
In addition to the chat bot foreseen for the e‑VISA pilot, eu‑LISA is involved in a number of projects that, amongst their components, foresee the use of Artificial Intelligence/ Machine Learning (ML). The implementation of AI-based solutions in the EU JHA domain requires large-scale efforts, and thus a collaborative approach and a flexible governance structure. The Agency is in favour of a phased approach, that does not imply huge investments from the beginning, and creating synergies between the different JHA domains (migration, asylum, law enforcement, etc.), thus avoiding a silo approach. This will also facilitate the possibility of capitalising on existing IT investments.
eu-LISA is clearly open to discussion and eager to focus on the changes, benefits and opportunities brought about by these technologies. Applying AI in a lawful, ethical and technologically robust manner is imperative. The Agency firmly believes in the creation of common AI capabilities that will allow internal and external synergies, as well as JHA domain communities to develop, test, and explore solutions for the future.
1. AI for the sBMS and biometric matching: Within the shared Biometric Matching Service (sBMS), the Agency is currently implementing a deep learning solution for biometric matching capabilities. The sBMS architecture was designed from the ground up with the utmost consideration for security aspects regarding access control and separation of data.
2. AI for the CRRS: The Common Repository for Reports and Statistics (CRRS), foreseen as part of the Interoperability architecture, will create centralised capabilities for reporting and analytics based on a central data repository and data lakes. There is a unique opportunity to exploit AI's potential to increase benefits from the CRRS deployment, as this allows:
3. AI Test Lab establishment The Agency will establish an AI Test Lab to facilitate the development and implementation of AI-based solutions in a way that is sustainable and with a view to further strengthening internal AI capabilities. The test lab could also support the testing of AI solutions by the Member States, in addition to eu‑LISA's own. This would be a welcome opportunity for the Agency and its stakeholders to further develop their learning curves and build on one another's experiences.
4. Impact assessment of hosting a data space for law enforcement In June 2020, eu‑LISA provided a first substantial contribution to a study on the requirements of data spaces for law enforcement purposes. A preliminary impact assessment on the creation and hosting of a centralised AI-enabled data space was carried out. eu‑LISA has recently continued the analysis with a view to extending the requested impact assessment to cover the option of a federated (against a fully centralised) solution. The analysis and discussion are ongoing.
eu-LISA prepares for engagement in the implementation of parts of the EU Framework Programme for Research and Innovation
With the entry into force of the revised establishing Regulation, the Agency's research mandate has been extended to include the opportunity to contribute to the implementation of parts of the EU Framework Programme for Research and Innovation. To put this into practice, the Agency has been collaborating with the Commission on preparing a working arrangement, defined as Terms of Reference. These Terms of Reference foresee the involvement of the Agency in the identification of priorities for research and innovation projects, in the areas related to the operational management of the large-scale IT systems entrusted to the Agency, as well as the possibility to engage with the projects throughout their lifecycle. We aim to sign the Terms of Reference with the Commission in Q1 of 2021. As a first step, eu-LISA has already supported the European Commission in the evaluation of research proposals submitted within the scope of the H2020 Security call for 2020 project proposals. We are very much looking forward to the opportunity to get involved in research and innovation activities aimed at making the Union safer and more secure.
EU Innovation Hub for Internal Security
In October 2019, Europol presented a proposal to establish an Innovation Hub at Europol. The EU Innovation Hub would act as an observatory for new technological developments, drive innovation, and ensure effective alignment and cooperation between relevant actors in the development of new technological solutions for internal security. Along with other JHA agencies, eu-LISA has been an active supporter of and contributor to this initiative from the very beginning. eu-LISA recently volunteered to support the development of the Hub by devising operational and financial frameworks, as well as proposing a methodology for a priority-setting process at the Hub. The Agency is committed to this initiative and will continue to support this important project in the coming months and years.
Innovation guides and enables us to keep working for...
A Safer Europe, 24/7
eu-LISA will present the latest EES developments at SMI's 14th Annual Border Security Conference, scheduled to take place online.
The TELEFI conference aims to facilitate information exchange on organisational, technical and legal aspects of facial recognition as used in criminal investigations.
eu-LISA's Executive Director was a panellist at the two online conferences organised by the German Presidency of the Council of the European Union.
It is now the 3rd time that the Board's meeting was held fully online. The Board adopted the Agency's Single Programming Document for 2021 to 2023, outlining all the programme- and project-based activities that eu-LISA is to perform to meet its strategic goals.
The Heads of JHA Agencies came together for a virtual annual meeting, where they had a first exchange on Artificial Intelligence and digital capacity building, training tools and innovative learning. "At eu-LISA we firmly believe in the creation of common Artificial Intelligence capabilities that will allow internal and external synergies, as well as JHA domain communities to develop, test, and explore relevant solutions for the future", said eu-LISA Executive Director Krum Garkov.
Common goals are the foundation for cooperation between eu-LISA and FRA for the benefit of citizens who gain from innovative technologies as much as from ensured respect for fundamental rights and data protection. The two Agencies signed a multiannual Cooperation Plan, which sets out activities that will provide benefits through joint actions.
As part of its cooperation strategy his the EU Council presidencies, eu‑LISA met up with the Portuguese team that will managed the upcoming presidency to provide expertise on topics related to the JHA IT architecture.
In his keynote speech, Krum Garkov, Executive Director of eu‑LISA, presented the Agency's involvement and contribution to the cybersecurity field.
Executive Director Krum Garkov presented eu‑LISA's angle on the ESS and the EU Smart Borders programme in a webinar organised by Vision-Box.
eu-LISA, Eurometropol, the Strasbourg City Magistrate and the French Ministry of the Interior formalised a land transfer that will host the extension of our Agency's operational site in Strasbourg.
The new Pact on Migration is of particular interest to eu‑LISA and its stakeholders as the current and future IT systems run by the Agency can have a positive impact and provide lasting support to Member States as they handle migration and asylum matters.
The two European Union Agencies have signed a multiannual Cooperation Plan, which sets out actions and activities that will enhance the cooperation and highlight the results of both parties from 2020 to 2022.
A delegation from the Estonian Information System Authority (RIA / Riigi Infosüsteemi Amet) and CERT-EE (Computer Emergency Response Team Estonia) visited the Agency's headquarters in Tallinn.