Vision and Ambition

Design, implement and validate a privacy-preserving law enforcement and decision-support platform to tackle critical challenges in cyber security, fully compliant with EU ethical standards for AI.

PRESERVE will leverage AI-based models to analyse large amounts of data from various sources and identify patterns linked to cybercrime and terrorism, enabling proactive threat detection and prevention.

The PRESERVE toolkit will serve as a critical asset for detecting and neutralizing emerging threats, while significantly advancing secure, cross-border collaboration among LEAs and establishing a regulation-compliant model for responsible AI use in crime detection and prevention.

Objectives

  • Develop a robust web-based big data analytics tool, compliant with privacy and data protection laws, supporting LEAs in their mission to detect and counter emerging threats. 
  • Provide strategies and solutions to reconcile big data analysis with EU data protection principles and regulations. 
  • Create solutions for gathering, normalising, correlating, and cross-referencing data across databases. 
  • Develop techniques to identify online behaviours related to cybercrime using web and social media analysis
  • Develop training curricula and programmes to train law enforcement personnel on big data analytics and responsible data use. 
  • Generate quality datasets for training and testing while ensuring lawful data sourcing and access. 
  • Validate the technical solutions developed within PRESERVE in real-world use cases. 
  • Maximise the disruptive potential of the PRESERVE’s solution through communication, dissemination and exploitation activities

PRESERVE Toolkit

The PRESERVE AI toolbox will rely on a variety of AI models, such as Natural Language Processing (NLP), Computer Vision (CV) and User and Entity Behaviour Analytics (UEBA) techniques. Federated Learning technologies will allow these AI models to be trained across decentralized systems without requiring data sharing between the different Law Enforcement Authorities (LEAs). Privacy Enhancing Technologies will ensure that sensitive information remains protected throughout the entire process, enabling LEAs to safeguard individual privacy and ensure unbiased and non-discriminatory policing.  

Once evidence is generated, the PRESERVE visualization tool will enable LEAs to enhance the investigation process through advanced, easy-to-use visual tools.  

LEAs will receive training on the PRESERVE tools through specially tailored, comprehensive curricula and programs designed to ensure effective tool usage. 

PRESERVE Technological Building Blocks 

Four Use Cases

The PRESERVE tool will be validated through four high impact use cases:

Detection of hate speech in online communication

Fighting terrorism recruitment and radicalisation by tracking social media and clandestine networks

Enhancing the investigation and prosecution of child sexual abuse through cyber monitoring and data cross-referencing

Identifying illegal drug trafficking