The rapid rise of synthetic media and deepfakes has introduced new, complex challenges for law enforcement, forensic experts, and judicial systems across Europe. Manipulated audio, video, and text can undermine investigations, distort evidence, and erode public trust. Addressing these risks requires not only advanced detection technologies, but also strong ethical, legal, and governance foundations.
This is where the connection between EITHOS and DETECTOR becomes relevant: EITHOS laid useful groundwork for trustworthy and ethical AI in law enforcement, and DETECTOR builds on this foundation to tackle the concrete challenge of deepfake and synthetic media detection.
Foundations for DETECTOR
EITHOS (European Identity THeft Observatory System) addressed the responsible and legitimate use of AI by Law Enforcement Authorities (LEAs) through cross-cutting ethical, legal, and governance frameworks that prioritised fundamental rights, accountability, transparency, and meaningful human oversight. Beyond technical considerations, the project adopted a strong societal and human-centred perspective, examining the social, psychological, and behavioural impacts of identity theft and AI-enabled crime on citizens and victims. Through extensive social research, citizen engagement, awareness and education activities, and the development of victim-support resources, EITHOS linked ethical and legal safeguards with public trust and social resilience, laying the groundwork for AI systems in law enforcement that are not only lawful and effective, but also socially acceptable and trusted.
Through close engagement with LEAs, security practitioners, and societal stakeholders, EITHOS translated these principles into practice by defining ethics-by-design and trust-by-design approaches for AI systems, exploring governance and oversight models grounded in real operational and legal constraints, and addressing risks related to bias, misuse, explainability, and evidentiary reliability. The project also delivered practical guidance, training materials, and operational recommendations that embedded European values into real-world law enforcement contexts.
DETECTOR’s Scope
Building directly on these conceptual and ethical foundations, DETECTOR (Deepfake Evidence and Technology for Forensic Content Oversight and Research) applies them to one of today’s most pressing AI-enabled threats: synthetic media and deepfakes. Specifically, DETECTOR focuses on:
- detecting manipulated and AI-generated audio, video, and textual content;
- supporting forensic institutes and LEAs in verifying the authenticity and evidentiary value of digital content;
- improving the reliability and legal robustness of deepfake detection methods; and
- embedding human oversight, transparency, and ethical safeguards into detection workflows.
Alongside technology development, DETECTOR places strong emphasis on training and capacity building, equipping LEAs and forensic experts with the knowledge needed to understand, interpret, and responsibly use deepfake detection outputs. In doing so, DETECTOR translates high-level ethical principles into operational tools, procedures, and skills that can be applied in real investigations and judicial processes.
The Team
Continuity between EITHOS and DETECTOR is also reflected in the composition of their consortia. In both projects, CERTH acts as the coordinating partner, ensuring strategic, scientific, and operational continuity. This is further reinforced by the participation of VICOMTECH and Engineering (ENG) in both initiatives, allowing knowledge, methodologies, and lessons learned to be transferred effectively from EITHOS to DETECTOR.
In EITHOS, close collaboration with three Law Enforcement Authority partners (the Hellenic Police, the Spanish Ministry of Interior, and the Swedish Police) helped ground the project’s ethical, technical, and societal work in real operational contexts. DETECTOR builds on this experience through collaboration with two Law Enforcement Authorities, the Portuguese Criminal Police (Portuguese Ministry of Justice, PJ) and the Finnish Police (SISAMINISTERIO, SM), as well as three forensic institutions: the Netherlands Forensic Institute (NFI), the Italian Ministry of Defence (RACIS), and the UK Home Office (HO). This strengthened mix of LEAs and forensic experts reinforces DETECTOR’s focus on evidentiary robustness, forensic validation, and real-world applicability in investigative and judicial contexts.
Validation
As with EITHOS, DETECTOR actively involves practitioners in requirements definition, tool testing and validation, and the assessment of legal, ethical, and operational implications. In DETECTOR, this engagement is further extended to Law Enforcement Authorities and forensic practitioners, who contribute directly to defining requirements for deepfake and synthetic media detection, validating forensic tools in investigative and judicial contexts, and assessing usability, explainability, and evidentiary value. This continuity ensures that detection technologies are fit for purpose, trusted by practitioners, and aligned with European legal and procedural standards.
Together, EITHOS and DETECTOR illustrate a clear research trajectory: EITHOS established trust, ethics, and governance for AI in law enforcement; DETECTOR applies these principles to the concrete challenge of deepfake and synthetic media detection. By building on EITHOS, DETECTOR strengthens Europe’s capacity to respond to the growing threat of synthetic media without sacrificing accountability, fundamental rights, or public trust, turning ethical AI principles into practical, forensic-grade solutions.
Stakeholders from law enforcement, forensic science, policymaking, and civil society are invited to join the DETECTOR Community to engage with project’s, activities, contribute their perspectives, and help shape the responsible use of deepfake detection technologies in real investigative and judicial contexts. To stay informed, subscribe to the newsletter.
