As internal security issues become increasingly pressing and the technologies to address them are refined with the help of developments in artificial intelligence, this rise in power raises crucial questions about fundamental rights, governance, and fairness. Many countries—particularly in Europe, which is the focus of this article—are striving to address these challenges by relying on hybrid AI that maintains the right level of human oversight and explainable AI that ensures the transparency and reliability of processes.
Towards “Smart Borders”…
In a few months, the European Union is expected to launch two integrated systems known as “Smart Borders”:
- the EES (Entry/Exit System) will record all entries and exits of third-country nationals into the Schengen area using fingerprints, facial recognition, and automatically recorded biometric identifiers, with integrated AI optimizing alerts in the event of abnormal situations (exceeding the length of stay, unusual movements, etc.).
- ETIAS, or European Travel Information and Authorization System, is an electronic authorization system for entry into the Schengen area of the European Union for visa-exempt travelers: AI promotes interoperability between member countries by cross-referencing available databases (Europol, Interpol, Schengen Information System), but also enables predictive risk analysis.
In fact, according to a report published last January (1), in 2024 Europe accounted for the largest share of the global market for automated border control (ABC), with 30% of a rapidly expanding market estimated at nearly €2 billion ($2.25 billion) and likely to triple within ten years.
The main uses of AI to enhance security, automate procedures, and improve migration management include:
- biometric identification, as mentioned above;
- automated document analysis systems that enable instant document verification and detection of inconsistencies;
- algorithmic risk assessment, i.e., analyzing data to identify suspicious behavior and/or potential risks;
- automated surveillance integrating the use of autonomous drones, increasingly intelligent sensors (acoustic, thermal, vibratory), and cameras to monitor border areas: a proven asset for monitoring hard-to-reach areas (seas, forests), embedded AI enables drones to detect movements, classify objects (humans, animals, vehicles) and transmit alerts in real time;
- Predicting migration flows using models developed from data fusion (drones, satellites, solar sensors, cameras, and social media), with multimodal AI also capable of combining video, text, and thermal signals in a single analytical interface.
- Some countries are also developing autonomous patrol robots, which have become increasingly sophisticated over the years. Some experts believe that these security robots are set to revolutionize not only border control, but also the management of sensitive and strategic infrastructure security (2).
While this automation facilitates procedures, the results of some pilot projects have nevertheless raised criticism and warnings about the risks of algorithmic errors or bias, and even human rights violations.
A report on the use of surveillance and artificial intelligence technologies in Spain (in Ceuta, Melilla, and the Canary Islands), entitled “Digital technologies for migration control at Spain’s southern borders” (3), highlights among the limitations observed a data management system that can lead to incorrect biometric records and use aimed primarily at speeding up returns. Better training for officers, harmonization of practices, and the inclusion of legal and ethical safeguards are among the report’s recommendations.
Similarly, the use of emotional AI in the iBorderCtrl project, which was tested in Hungary, Greece, and Latvia between 2016 and 2019, was controversial. This sought to identify risk profiles by detecting suspicious or deceptive behavior through the analysis of micro-facial expressions, body language, and responses to stress. The experiment was criticized for its lack of reliability and denounced by NGOs concerned about cultural bias.
The issue of transparency highlighted in this case study, which is emblematic of the tensions between technological innovation and migration policy, regularly arises in assessments of the use of AI in the field of security, as the algorithmic methods used are partially opaque and fuel criticism about the lack of explainability. This has led to the development of what is known as explainable AI or “XAI” – or rather its refinement, since the origins of explainable or interpretable AI date back to work carried out by the US Defense Advanced Research Projects Agency (DARPA) in 2016.
(To Be Continued >>>See Part II)
(By Murielle Delaporte)
Notes & References
(1) See: https://www.fundamentalbusinessinsights.com/fr/industry-report/automated-border-control-market-8217
(2) See for instance: >>> https://www.runningbrainsrobotics.com/fr/6-tendances-securite-et-robotique-2025/
(3) See: https://euromedrights.org/fr/publication/nouveau-rapport-une-analyse-des-technologies-de-surveillance-numerique-pour-le-controle-des-migrations-aux-frontieres-sud-de-lespagne/
Photo © European Parliamentary Research Service, Securing EU borders with artificial intelligence, Youtube video, 01/04/2022