Ethical and Privacy-preserving Big Data platform for Supporting criminal investigations

Project Information

PRESERVE

Grant agreement ID: 101168309

DOI

10.3030/101168309

EC signature date 22 July 2024

Start date 1 September 2024

End date 31 August 2027

Funded under

• Civil Security for Society

Total cost

€ 6 365 398,54

EU contribution

€ 5 338 197,00

Coordinated by

FUNDACION CENTRO TECNOLOXICO DE TELECOMUNICACIONS DE GALICIA
Spain

Periodic Reporting for period 1 - PRESERVE (Ethical and Privacy-preserving Big Data platform for Supporting criminal investigations)

Reporting period: 2024-09-01 to 2026-02-28

Summary of the context and overall objectives of the project

In the context of the EU’s objective to enhance security while protecting fundamental rights, PRESERVE addresses the growing complexity and digitalisation of crime, including cybercrime and terrorism. Law Enforcement Authorities (LEAs) face increasing volumes of heterogeneous data from social media and the surface, deep, and dark web, which exceed traditional investigative capacities and challenge cross-border cooperation. At the same time, the use of Big Data and AI is limited by strict regulatory frameworks and concerns related to privacy, bias, and transparency.

PRESERVE tackles these challenges by developing a privacy-preserving, AI-driven big data analytics platform that enables secure and collaborative intelligence sharing across Member States, leveraging technologies such as federated learning. The platform integrates advanced capabilities—including Natural Language Processing, computer vision, graph analysis, and behavioural analytics—to improve threat detection and investigation processes.

The project adopts an interdisciplinary approach by incorporating Social Sciences and Humanities (SSH), ensuring legal compliance, ethical alignment, and stakeholder engagement. Overall, PRESERVE aims to deliver technological tools, training, and policy guidance to strengthen LEAs’ capabilities, foster cooperation, and support trustworthy AI adoption in Europe.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

During the first reporting period, the PRESERVE project has progressed from initial design to the development and early integration of its core technical components.

Key achievements include the definition of a robust data management framework aligned with FAIR and GDPR principles, enabling consistent data handling and preparation for integration across the platform. A comprehensive set of end-user requirements was identified and validated, providing a clear technical baseline for system design and tool development.

On the data layer, significant progress was made in data collection and processing, including the development of the MEDUSA crawling tool for gathering textual and visual data from online sources. In parallel, data quality guidelines and interoperable data exchange models and APIs were designed to support secure and efficient data sharing.

The project also advanced the development of AI-based analytical components, including NLP tools for cybercrime analysis, computer vision models for image and video processing, and behaviour analysis models. These developments were supported by work on federated learning and synthetic data generation to address data privacy and availability challenges, as well as initial steps toward explainable and fair AI.

In terms of system design, a reference architecture was defined, incorporating privacy-preserving mechanisms and interoperability requirements. Initial integration activities were launched, including the setup of a secure infrastructure, deployment environments, and continuous integration and testing pipelines.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

By developing an innovative AI-based, privacy-preserving decision-support toolbox for European Law Enforcement Authorities (LEAs) that combines proactive threat detection and strong compliance with EU fundamental rights and regulatory requirements, PRESERVE lays the foundation for a transformative, regulation-compliant European security solution that advances beyond current technological, ethical, and operational state of the art.

The combination of NLP, Computer Vision, and User and Entity Behaviour Analytics, Federated Learning, and Privacy-Enhancing Technologies, PRESERVE in one toolbox enables enabling secure, cross-border collaboration without compromising sensitive data or individual privacy that goes beyond traditional tools for LEAs.

The project’s expected impacts include a significantly enhanced operational efficiency for LEAs, improved capability to detect and prevent cyber-enabled crime and terrorism, improved protection of fundamental rights, and increased trust in the use of AI for security purposes. PRESERVE also contributes to European technological sovereignty by driving innovation in trustworthy AI, cybersecurity, and privacy-preserving technologies aligned with EU strategic priorities.

Extensive testing with end-users, demonstration in real-life operational environments, and further validation across all targeted use cases, are key for PRESERVE’s successful implementation and long-term uptake. Moreover, training and upskilling of existing teams at the LEAs will be critical for the successful implementation and deployment of the tools in the daily life of the LEAs. Further key requirements are the establishment of clear and supportive legal and regulatory conditions for the deployment of PRESERVE tools. This includes supportive national and EU policy frameworks, access to faster approval or validation pathways, and the availability of sandbox environments and relevant data for testing and refinement. The broader impact of PRESERVE also depends on recognised interoperability standards that strengthen trust, support market acceptance, and facilitate the expansion of the end-user community. The successful uptake of PRESERVE innovations by LEAs will require procurement frameworks reducing barriers to implementation that are innovation-friendly and supportive of EU-developed technologies.