BorderForce is a EU-funded Horizon Research Project whose aim is to develop AI-assisted green border surveillance, in order to control irregular border crossings of goods and people on the move. The project is currently developing technologies that will have real-time surveillance capabilities through the deployment of AI. These technologies will thus ensure “frequent monitoring and early threat assessment.” The objective of BorderForce is to create technologies that assist border guards in order to “improve regional stability and security” through constant monitoring and surveillance.
The initiative is also described as a project that will achieve full time and cost optimization as well as a high level of efficiency, thanks to “seamless” monitoring operations of people, goods and information. This will enable the authorities to address “challenges like migration, smuggling, and geopolitical tensions, particularly in crises.” The project will develop a series of new surveillance technologies that will include:
“Self-sufficient, transportable Command and Control Stations integrating AI based detection for autonomous optical edge sensing and (Anti-) UAV functionalities”: mobile operational units equipped with autonomous power supplies capable of automatically detecting, tracking and classifying individuals, vehicles, animals, vessels and UAVs directly on site without sending all data to a central server or requiring constant human operator intervention;
Surveillance towers, integrating data from autonomous monitoring sensors and UAV systems;
Satellite resources (Copernicus satellites, Cubesats) to gather crucial data;
“Drone swarm sensing with sophisticated flight plan updates”: multiple drones flying simultaneously in an interconnected network to scan vast areas of territory in a fraction of the time a single drone would require. This guarantees continuous coverage through algorithmic route management that dynamically adjusts flight paths upon detecting anomalies, thereby minimizing human intervention;
Open Source Intelligence (OSINT) platform to enable early threat assessment: a platform that collects and analyzes data from open and public sources (social media, online articles, blogs) to identify false online content, verify media authenticity, detect anomalies, and track physical threats starting from digital data. Features include “AI image detection, image manipulation detection, fake face detection, morphed face detection, deep fake detection,” and even “geolocation estimation,” allowing authorities to “estimate GPS coordinates from image data.”;
“Autonomous UAV-aided mesh wireless communication network with RPAS management (VTX Mesh),” designed to ensure continuous, stable, and high-speed network coverage even in the most remote and signal-deprived border areas;
A data fusion system, where “data fusion will be performed on features extracted from sensor data and the OSINT sub-system.” This means that the data collected from all deployed surveillance technologies (drones, sensors, OSINT platform, etc.) will be cross-referenced and processed through predictive analytics to classify potential risk by type and severity, ultimately generating threat evaluations and maps compliant with Frontex standards.
On the BorderForce website, it is stated that “ethical considerations and fundamental rights” will be prioritized. However, it remains vague as to which fundamental rights are being referred to and how they will be effectively guaranteed in practice. Equally concerning is BorderForce’s objective “to achieve a commonly accepted framework and system for internal and external border surveillance.” This explicit inclusion of internal borders suggests that these intrusive technologies are not intended solely for the EU's external frontiers. Instead, they could easily be deployed to monitor borders between EU Member States during temporary suspensions of the Schengen Agreement.
Another key operational dimension of BorderForce involves the premise upon which the project was conceived: bridging the information gap between digital intelligence from social media and real-time physical activity on the ground. Thus, the overarching goal is to develop technologies capable of scraping and scanning social media platforms to extract data aimed at preventing or suppressing border-related activities. This raises significant concerns regarding privacy rights and the potential for mass surveillance.
AlgorithmWatch reports that when it requested access to documentation regarding the project, authorities responded by heavily redacting detailed descriptions and objectives, under the pretext that this constituted “security-related information” forming part of “confidential strategies,” and that its disclosure would “undermine public security" or be "misused by malicious actors.” This lack of transparency represents a major barrier to the systematic mapping of such technologies, as States frequently invoke security concerns and the imperative of national sovereign control over borders to justify withholding information from the public. Additionally, this opacity is particularly troubling given that border guards played a pivotal role in driving the development of BorderForce, shaping it entirely around their own self-defined needs and operational use cases.
Finally, the project’s compliance with the EU AI Act warrants close and ongoing monitoring, particularly as BorderForce operates within the scope of high-risk AI systems for border surveillance. Under the AI Act framework, high-risk systems require continuous risk management processes as well as robust data governance to mitigate biased outcomes. Crucially, both the AI Act and the General Data Protection Regulation (GDPR) establish that AI should serve as a decision-support system rather than a fully autonomous decision-maker, underlining the necessity for meaningful human oversight to be maintained in final decision-making processes.
EU AI Act; General Data Protection Regulation (GDPR)