AI4SafeEurope Cluster: New Milestones, Shared Mission 

Guest Article by AI4SafeEurope Cluster led by Sister Project AVALANCHE 

Through the AI4SafeEurope Cluster, eight complementary Horizon Europe projects working at the intersection of AI, security, and law enforcement join forces leveraging AI to strengthen civil security. The cluster, which is coordinated by the project AVALANCHE, fosters joint discussions on key challenges and opportunities, including disinformation, cybercrime, border protection, interoperability, trustworthy AI, and other critical issues shaping the future of the sector. 

Since its launch in Spring 2025, the Cluster has already established an impressive track record. Through a growing portfolio of joint activities, it has strengthened collaboration between the participating projects, promoted the exchange of expertise and thus contributed to advancing a shared European vision for the responsible use of AI in civil security. 

Beyond joint workshops, webinars, and a shared Cluster newsletter, the AI4SafeEurope Cluster is fostering closer engagement with the broader community of Law Enforcement Authorities (LEAs) involved across its projects. This creates valuable opportunities to showcase project results, gather feedback from end users, assess the operational relevance of emerging AI solutions, and strengthen dissemination through a shared LEA-focused ecosystem. 

A recent milestone in this regard was the AI4SafeEurope Cluster Online Workshop for Law Enforcement Authorities in May, which brought together the participating Cluster projects and representatives from European LEAs. During the workshop, the cluster projects presented the solutions they are developing, followed by a discussion on operational needs and opportunities for future collaboration. The event was followed up with targeted efforts for cross-cluster matchmaking between cluster projects and participating LEAs. 

Updates from the participating projects: 

  • AVALANCHE has now entered its very active final phase, with integration, piloting and validation activities progressing in parallel. Following the Athens plenary meeting last month, partners reviewed technical progress and tested the 2nd platform release with two LEA end-users, gathering feedback to support final refinements. The 3rd release is currently underway, and work is focused on consolidating and assessing technical KPIs and preparing the final release of the AVALANCHE platform. 
  • CEASEFIRE, which was successfully completed in September 2025, delivered a highly innovative, holistic, multidisciplinary, high-tech, and versatile approach that significantly enhances and broadens the operational capabilities of EU Law Enforcement Agencies in their efforts to detect, analyse, and track cross-border illicit firearms trafficking activities. 
  • ENSEMBLE: Following its 3rd Plenary Meeting and first operational pilot in Nancy at the beginning of July, the project has entered a critical validation phase. Investigators and technical partners tested newly developed AI-driven tools against cross-border networks, cross-chain crypto flows, and dark web threats in simulated environments. Feedback from this milestone is currently driving tool refinements as the consortium maps out the upcoming deployment and evaluation phases.
  • FERMI was successfully completed in September 2025. However, project findings and key insights continue to be disseminated. Researchers from the Bavarian Police College’s Centre of Excellence in Police and Security Research introduced the ramifications of disinformation to an audience of approximately 180 students and 20 lecturers during the Digital Day of the East Bavarian Technical University on 26 June 2026. On 29 June 2026, insights from the FERMI project were presented at a major seminar on organised crime at the Bavarian Police Training Centre. 
  • GANNDALF has entered a new phase focused on refining its operational scenarios and preparing the upcoming pilots. LEA and technical partners are working closely to fine-tune scenarios, deployment requirements and evaluation activities, ensuring that AI-powered tools are tested under realistic operational conditions. The pilots will cover financial and data-related crime, secure collaboration, cyber protection of critical emergency infrastructure, and citizen cyber hygiene. 
  • NOTIONES: Over the last months, NOTIONES has focused on strengthening its impact and sustainability within the European security community. A key milestone was the NOTIONES Final Conference, held in Lille, France, in April. The event showcased key results and celebrated five years of collaboration, technology monitoring, and intelligence analysis. In June, NOTIONES participated in the TASM Conference 2026, sharing its work and gathering insights on AI-enabled extremism, online narratives, and OSINT. As NOTIONES nears completion, partners are finalising activities to ensure the project’s legacy and support future collaboration and innovation. 
  • PRESERVE (AI) is developing an AI-powered decision-support toolbox for cybercrime prevention and detection, fully aligned with EU ethical and legal standards. Recent highlights include the publication of a policy brief on AI-supported cybercrime investigations following the “Networking within the Security Research Community!” event initiated by PRESERVE and held in Madrid; the organisation of the third edition of the same event, held in Bari in April, which kicked off the drafting of a joint position paper; a successful mid-term review; and the successful launch of the project’s first pilot operations with its three partner law enforcement agencies in June. 
  • PRESERVE C-UAS successfully deployed its multimodal fusion engine prototype, connecting ODIN (online threat monitoring) with SECURE (physical sensor arsenal). Leveraging a database of 210 drone models, the system delivered highly encouraging detection and classification results. Operationally, PRESERVE’s initial standardized procedures were shared for the upcoming testing in two Pilot Use Cases. Concurrently, the inaugural training session certified 13 police agents, achieving an 88% user satisfaction rate. Finally, PRESERVE successfully presented its work during the first project review, with satisfactory results as well. Make sure to keep an eye on our channels for upcoming events after the summer break! 

Read more news about the A4SafeEurope Cluster in our news archive