How AI can support the European Border and Coast Guard

Conceptual representation of the protection and interaction of the personal data of the inhabitants of the planet and industrial data using artificial intelligence and the Internet of things, photo by kosssmosss/Adobe Stock

kosssmosss/Adobe Stock

Three baseline gaps should be addressed to maximise the opportunities provided by artificial intelligence in European border security: knowledge gaps between stakeholder groups; organisational, structural, cultural and expertise gaps; and broader gaps in the evidence base.

What is the issue?

The EU faces a number of threats at the external borders, stemming from factors such as irregular migration, people smuggling, human trafficking, illegal weapons, and drugs. Artificial Intelligence (AI) yields great potential to significantly enhance the effectiveness of security and operational activities at the borders and beyond, with direct and multiplying impact on internal security.

However, various technological and non-technological barriers might influence how AI materialises in the performance of border security functions. Frontex, the European Border and Coast Guard Agency, is tasked to proactively monitor and contribute to research and innovation activities relevant for border control and to keep the European Parliament, Member States, and the European Commission informed of new developments.

How did we help?

RAND Europe was commissioned by Frontex to explore the ways in which the European Border and Coast Guard can maximise the opportunity provided by AI to support the management of the European Union’s (EU) external borders. The study also examined how Frontex can support the EU’s wider goals in the development and deployment of cutting-edge, ethical and secure AI.

The research team collected data through case studies (via interviews and document review), and horizon scanning, before conducting a STREAM (Systematic Technology Reconnaissance, Evaluation and Adoption Method) workshop to comparatively assess the potential impact and feasibility of implementation of different AI-based capabilities. They also developed technology roadmaps for key AI technologies via interviews with technology experts and border security end-users, internal workshops and desk-based research to elaborate on the steps required to integrate AI-based systems into border security.

Research questions:

  1. What is the current landscape in the application of AI to border security?
  2. Which new and emerging AI-based systems may be applied to border security?
  3. In which areas of border security may new and emerging AI-based systems be applied?
  4. What steps are required to integrate AI-based systems into border security?

What did we find?

Current and potential future uses of AI cut across several border securities functions, including: situation awareness and assessment; information management; communication; detection, identification and authentication; and training and exercise. The AI systems currently used in development for border security purposes include ‘front-end’ capabilities (e.g. security gates and surveillance systems) and ‘back-end’ capabilities (e.g. automated machine learning).

The research team explored the opportunities, requirements and barriers for adoption for nine specific AI technology areas.

  • The nine technology areas assessed were: automated border control, maritime domain awareness, machine learning optimisation, surveillance towers, heterogeneous robotic systems, sUAS, predictive asset maintenance, object recognition, geospatial data analytics.
  • Initial assessment at the STREAM expert workshop indicated that all AI technology areas were considered to bring at least a moderate improvement to the ability of the European Border and Coast Guard to perform border security functions.
  • None were perceived to face overwhelming barriers to adoption.

Enablers and barriers to adoption of AI-based systems

Various barriers could pose significant challenges to end users and efforts to integrate AI-based systems in support of border security functions. These include:

  • Technological barriers e.g. algorithmic biases, cybersecurity vulnerabilities
  • Cost and commercial barriers
  • Insufficient understanding and awareness of AI among end users
  • Skills shortage and lack of expertise and organisational capacity
  • Lack of access to relevant technologies and dominance of non-EU technology suppliers
  • Ethics, human rights and regulatory barriers

Key enablers for the adoption of AI-based technologies include:

  • Advances in AI methods and adjacent technologies and iterative development of AI-based capabilities
  • Improvements in usability, e.g. simplification of technology interfaces
  • Commercialisation and democratisation of AI as enablers for decreasing costs
  • EU initiatives on wider AI research and development
  • Increasing public awareness and acceptance of AI-based technologies

What do we recommend?

The outputs of this study and the nine technology adoption roadmaps seek to provide a high-level overview of the main opportunities, challenges and requirements for the adoption of AI-based capabilities in European border security. Frontex could consider:

  • Working to address three overarching baseline gaps, namely: knowledge gaps between stakeholder groups; organisational, structural, cultural and expertise gap; and gaps in the evidence base.
  • Defining what role Frontex as an agency could play in shaping the future landscape of AI-based capabilities. The study outlines five possible roles for Frontex.
  • Identifying options for addressing key organisational structural and cultural barriers – including procedural and organisational inefficiencies and further assessing gaps in skills, expertise and resources.